James VI and Witchcraft: North Berwick, Daemonologie and Doubt

In 1597 the king’s printer, Robert Waldegrave, issued a short book in the form of a dialogue, signed “IAMES Rx”. Its author was the king of Scots, and its subject was witchcraft: whether it existed, what witches could do, and how they should be tried and punished. Daemonologie was not an armchair exercise. Six years earlier James had sat in on the interrogations of men and women accused of plotting his death by sorcery, and several of them were executed. Then, after he became king of England in 1603, the same monarch acquired a reputation for catching people who faked bewitchment.¹²³

These episodes are easy to compress into a tidy story of a zealot who mellowed. The evidence is less tidy, and it concerns real people, most of them women, who were tortured on the strength of beliefs they could not disprove.

Storms, a royal marriage and a servant in Tranent

In 1589 James sailed to Scandinavia to fetch his bride, Anne of Denmark, after storms had driven her own ship back. Bad weather dogged the couple’s crossings, and a vessel carrying wedding gifts was lost. In Denmark, officials tried women accused of raising those storms by sorcery, a type of charge familiar there.⁴

Scotland already had a law that made witchcraft a capital crime. The act of June 4, 1563, forbade anyone “of quhatsumever estate, degre or conditioun” to practise witchcraft, sorcery or necromancy, or to consult those who did, “under the pane of deid”.⁵ What changed in 1590 was the scale of the accusation and the presence of the king.

The case began in Tranent, east of Edinburgh. According to the pamphlet Newes from Scotland, David Seaton, a deputy bailie, grew suspicious of his servant Geillis Duncan, who had begun to cure the sick. With the help of others he tortured her with thumbscrews (“Pilliwinckes”) and by wrenching her head with a cord. Getting no confession, they searched her body for the Devil’s mark, claimed to find it on her throat, and only then did she confess. In prison she named others.² Among those named was Agnes Sampson, a healer and midwife whose clients, the Survey of Scottish Witchcraft notes, ranged from poor households to lairds and their wives. The Survey’s record of her case opens on October 22, 1590, and lists James VI among the investigators.⁶

The North Berwick confessions

The confessions that followed grew into a story of a conspiracy. In the version that reached print, witches sailed in sieves to the kirk of North Berwick on Halloween, danced to a Jew’s harp played by Geillis Duncan, and heard the Devil preach from the pulpit. Asked why he hated the king, the Devil replied that “the King is the greatest enemy he hath in the worlde”. Witches were said to have christened a cat, bound parts of a corpse to it, and thrown it into the sea off Leith to raise the storm that sank the boat carrying the queen’s gifts.² The court records summarized by the Survey contain similar material, including a cat cast into the sea to raise a wind.⁶⁷

The pamphlet shows James as an engaged and skeptical listener. Agnes Sampson, it says, confessed things so “miraculous and strange” that “his Maiestie saide they were all extreame lyars”. She then took him aside and repeated words he had exchanged with his queen on their wedding night in Norway, and the pamphlet says he swore that “all the Diuels in hell could not haue discouered the same”.² The passage was written to show that the king’s belief had been earned, but the same text records how belief was produced. Sampson was examined at Holyroodhouse and denied everything. She was then shaved, her head was wrenched with a rope for nearly an hour, and she confessed once a mark was found on her body.² An English ambassador’s report, as summarized by the Survey, said she admitted 58 of the 102 charges in her indictment.⁶

John Fian, a schoolmaster at Prestonpans, suffered most in the printed account. Tortured with the boots, he confessed and signed his confession before the king, then escaped, was recaptured, and denied it all. James, the pamphlet says, imagined that he had made a new league with the Devil and been marked afresh, though a search found no mark. Fian’s fingernails were torn out with pincers and needles pushed beneath them, and his legs were crushed in the boots. He kept to his denial, saying that what he had said before “was onely done and sayde for feare of paynes which he had endured.” He was strangled and burned on the Castle Hill of Edinburgh at the end of January 1591.²

Historians read the pamphlet with care. Its authorship is uncertain, and Christina Larner argued that it was aimed at an English public, at a time when James hoped to inherit Elizabeth’s throne.⁴⁸ It also gives several key confessions to an “Agnis Tompson” of Edinburgh, while the court summaries place the fullest North Berwick narrative in Sampson’s case.²⁶ The Edinburgh survey team are blunt about its tone: the pamphlet “describes these tortures with relish. But they were unusual.” The most common form of torture in Scotland, they note, was sleep deprivation; physical tortures of this kind were used only occasionally.⁹

Treason, a rival earl and a jury on trial

The affair became political. Geillis Duncan’s confessions named Francis Stewart, Earl of Bothwell, the king’s cousin and a long-standing rival, though she said she had only heard him and never saw him.⁷ In April 1591 James ordered Bothwell’s arrest for conspiring his death by witchcraft; Bothwell escaped and was later forfeited by Parliament.⁴

Two of the accused stood out in the pamphlet because, it said, they had been “reputed for as ciuill honest women as any that dwelled within the Citie of Edenbrough”.² Barbara Napier was acquitted by her assize, the Scottish jury, of making wax images of the king and of attending the North Berwick meeting. The crown then prosecuted the assize itself for wilful error, and James ordered doctors to examine whether Napier was pregnant.¹⁰ Euphame MacCalzean, who held the lands of Cliftonhall near Edinburgh, was accused of trying to destroy the king and queen and of harming her own relatives. A letter from James to his minister Maitland concerned her torture, and she was tried in June 1591. In 1592 an act restored Cliftonhall to her heirs.¹¹

How many people were caught up in all this is uncertain. The Survey counts 3,837 people accused of witchcraft in Scotland across the whole period of the 1563 act, 84 percent of them women, and identifies 1590-91 as the first of five intense panics. Its compilers found no evidence of any organized witch cult.⁹

Daemonologie: a king answers the doubters

James explained his purpose in the preface. “The fearefull aboundinge at this time in this countrie, of these detestable slaues of the Deuill, the Witches or enchaunters,” had moved him to write. His targets were two skeptics: the Englishman Reginald Scot, who, James complained, “is not ashamed in publike print to deny, that ther can be such a thing as Witch-craft”, and the physician Johann Weyer, whom James called “a German Phisition” and whose defense of accused witches, James said, revealed him as one of them.¹

The book takes the form of a conversation between Philomathes, who doubts, and Epistemon, who explains. When Philomathes asks why there are “twentie women giuen to that craft, where ther is one man”, Epistemon answers that “that sexe is frailer then man is” and so easier for the Devil to trap.¹ On punishment the text is unbending. Witches “ought to be put to death according to the Law of God, the ciuill and imperial law, and municipall law of all Christian nations”, and those who consult them are as guilty as the practitioners. A ruler who spares them sins as Saul did in sparing Agag.¹

Yet the dialogue also contains a caution that James would return to. “Iudges ought indeede to beware whome they condemne”, Epistemon says, since condemning the innocent is as great a crime as letting the guilty go. The safeguard is weak, though: in the same breath he allows the testimony of other accused witches, reasoning that only witches can witness the deeds of witches. He also endorses two tests, pricking for the Devil’s insensible mark and “fleeting on the water”, claiming that “the water shal refuse to receiue them in her bosom, that haue shaken off them the sacred Water of Baptisme”.¹

The year of publication was itself a year of panic. In 1597 a woman named Margaret Aitken claimed she could identify witches by sight and was taken from town to town for that purpose, until she was found to be accusing and clearing the same people. The swimming test was used on an unknown number of suspects that year and then, the Survey reports, seems to have been discredited in Scotland.⁹⁴ The national commissions for trying witches were recalled later in 1597.⁴

England, 1604: a new statute

When James came south in 1603, his English Parliament soon replaced Elizabeth’s witchcraft act of 1563. The new statute of 1604 made it a felony punishable by death to invoke, consult, “entertaine, imploy, feed, or reward any evil and wicked spirit,” to take a corpse or any part of one for use in witchcraft, or to kill or lame anyone by witchcraft. Lesser offenses, such as using magic to find treasure or lost goods, to provoke “unlawful love,” or to harm cattle, brought a year in prison and the pillory on a first conviction and death on a second.¹² Whether James drafted or pressed for it is not established by the sources used here; what is clear is that it widened the range of capital offenses.

The king who looked for fakes

At the same moment, as the legal historian Krista Kesselring observes, English authorities were also explaining away claims of possession and witchcraft as illness or imposture, and Archbishop Richard Bancroft was campaigning against exorcism.¹³ James took part. In 1604 Anne Gunter, a young woman in North Moreton in Berkshire, suffered fits, vomited pins, and accused three women of bewitching her; two were tried at Abingdon in March 1605 and acquitted. Her father then brought her to the king in Oxford in August 1605. James passed her to Bancroft, whose chaplain Samuel Harsnett soon had her confessing that she had faked the symptoms at her father’s direction, partly by copying a printed account of an earlier case. James himself wrote what the historian James Sharpe calls “virtually a psychiatric report” on her to his chief minister.³ The Star Chamber case that followed has no surviving verdict.³

Sharpe sums up the pattern: “James, in England at least, was as willing to demonstrate his expertise in matters of witchcraft by exposing fraudulent cases as by finding groups of malefic witches.”³ Leicester in 1616 has been offered as a further example. Letters of Alderman Robert Heyrick record that nine women were to be hanged in July for bewitching a boy of 12 or 13 from Husbands Bosworth, whose fits had been shown to the judges, and that in October five more accused women were set free and a sixth had died in gaol.¹⁴ Heyrick’s letters do not mention the king. A local historian, writing much later, says it “seems” James’s intervention lay behind the release and that he judged the boy a fraud; that attribution should be treated as plausible rather than proven.¹⁵

A changed mind, or the same mind?

The historian Brian Levack has argued that James did change, turning over time from a hunter of witches into a doubter, and that argument has reshaped how his role is assessed.⁴ The texts also allow a reading of greater continuity. Daemonologie already warned judges against convicting the innocent, even while it demanded death for the guilty and accepted accomplices’ accusations and the swimming test.¹ In England, exposing a fraud was another way for a learned king to show he knew the difference. And the statute passed early in his English reign had made more kinds of witchcraft punishable by death.¹²³¹³

Scholars still weigh how much responsibility for the Scottish prosecutions belongs to James himself. Larner saw his personal role in 1590-91 as central and doubted claims that he joined the investigation late; Levack has argued that he did come late, after suspects had already confessed to a plot against him. Levack and Julian Goodare have also pointed to Scotland’s legal system, and Goodare to local elites, as the main drivers of its high rate of prosecution.⁴⁸ Lawrence Normand and Gareth Roberts have edited Daemonologie together with the North Berwick documents.¹⁶

What the records establish beyond dispute is narrower and heavier. Geillis Duncan, Agnes Sampson, John Fian and others were tortured into confessing, several of them were examined in front of their king, and their answers were then printed as proof of a plot. As the Edinburgh survey puts it, those convicted in Scotland were almost always strangled at the stake and their bodies burned.⁹ Fian’s last recorded words in the pamphlet, that he had confessed only out of fear of pain, are a plain reminder of how the evidence was made.²

Footnotes

1. James VI, Daemonologie (1597), Preface; Book 2, ch. 5; Book 3, ch. 6 (Gutenberg #25929). ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶

2. Newes from Scotland (London, 1591?), printed in Gutenberg #25929. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹

3. Sharpe, Times Higher Education, November 5, 1999. ↩ ↩² ↩³ ↩⁴ ↩⁵

4. Zacharia 2020, Tufts honors thesis, introduction and chs. 2-3. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷

5. Goodare 2005, Church History 74(1), opening quotation of the 1563 act. ↩

6. Survey of Scottish Witchcraft, case C/EGD/63 (Agnes Sampsoune). ↩ ↩² ↩³ ↩⁴

7. Survey of Scottish Witchcraft, case C/EGD/110 (Gelie Duncan). ↩ ↩²

8. Larner 1973 (catalog record only). Larner’s views are given as summarized in Zacharia 2020, which cites her Enemies of God (1981) and Witchcraft and Religion (1984). ↩ ↩²

9. Survey of Scottish Witchcraft, “Introduction” (questions and answers). ↩ ↩² ↩³ ↩⁴

10. Survey of Scottish Witchcraft, case C/EGD/103 (Barbara Napier). ↩

11. Survey of Scottish Witchcraft, case C/EGD/104 (Ewfame Makcalzene). ↩

12. The National Archives, 1604 Act Against Witchcraft (C 65/181), transcript. ↩ ↩²

13. Kesselring, Legal History Miscellany, October 31, 2023. ↩ ↩²

14. Nichols, History and Antiquities of the County of Leicester, vol. 2, Husbands Bosworth (Heyrick letters, 1616). ↩

15. Marquis, LAHS blog, June 9, 2024. ↩

16. Normand and Roberts 2000 (catalog record). ↩