The girl who barked like a dog — Boston, 1688

In November 1688 an Irish woman was hanged in Boston for bewitching four of the Goodwin children. Her name was Ann Glover. She was a Catholic, she was poor, and by every account she was either unwell or unable to defend herself in the language the court was using. She is remembered now as the last person executed for witchcraft in that city.

She is not the interesting part of the story. The interesting part is the child who put her there.

The eldest of the Goodwin children was about thirteen. The most experienced physicians in Boston examined the family and pronounced the children bewitched, which is what a physician did in 1688 when the symptoms did not fit anything he had a name for. What followed was documented at length by Cotton Mather, the minister who took one of the girls into his own household to see whether he could cure her.

His account is not the account of a credulous fool. It is the account of a clever man being played, in writing, at length, by a child, and never once noticing.

What the children did

By Mather's description they barked like dogs, then purred like cats. They flew.

"They would fly like geese, and be carried with an incredible swiftness, having but just their toes now and then upon the ground, sometimes not once in twenty feet, and their arms waved like the wings of a bird."

One of them went further than the others. She complained that the old Irish woman had tied an invisible noose around her neck and was choking her, and her face would change colour and distort exactly as a choking person's does. She said an invisible chain was fastened to her leg, and limped about as though it were. She said she was in an oven, and sweated and flushed and looked roasted; then she cried out that cold water had been thrown on her, and her whole body shook. She claimed the devil came to her in the shape of an invisible horse, and cantered and galloped and trotted and ambled around the rooms and the entries, and went up the stairs with the exact toss and bound of somebody riding.

That is not a description of possession. It is a description of a very good actress, working an audience that had already decided what it was watching.

Three languages and a fourth

Mather wanted to test her. Wanting to say something in her presence that she should not understand, he spoke in Latin. She guessed what he had said, and he was amazed. He tried Greek, and she was equally successful. He tried Hebrew, and she caught the meaning instantly.

Then he tried the language of the people whose land the town was built on, and she pretended not to understand a word.

Mather's conclusion, written down in earnest, was that the evil spirit she was in league with was fluent in Latin, Greek and Hebrew, but not in the Indian tongue. It did not occur to him that a minister's household is a place where a bright child can pick up Latin, Greek and Hebrew from the air, and that the one language nobody in that house taught her was the one she had no reason to know. The evidence fitted his world exactly, which is why he never questioned it.

The books she could not read

Then he handed her books, and she gave him his own opinions back, one at a time.

A book by a Quaker, a sect Mather violently opposed: she read it off with ease and pleasure. A book written against the Quakers: she could not read a word of it. Books by Catholics: fine. The Assembly's Catechism: not a syllable. The Book of Common Prayer, which Mather's own church had rejected and which she took to calling her Bible, would relieve her suffering the moment it was put into her hands. She could read the scriptures easily in the Prayer Book, but not in the Bible. She could not read the Lord's Prayer in either.

Every one of those results maps onto a prejudice of the man holding the book.

That last detail is where the whole thing falls apart, and Upham, writing two centuries later, saw it plainly. A girl who could read Latin and Greek and Hebrew, but who could not read the Lord's Prayer in her favourite volume, was not demonstrating a devil's learning. She was demonstrating that she had never been taught to read fluently at all, and was reciting what she had heard and refusing what she had not.

The room the demons would not enter

The performance got better as the audience got more important. Mather's study was, in her account, a place the demons could not follow her into. Dragged up the stairs towards it she became so heavy that it took several people to move her, screaming that they had told her she must not go in — and the moment she was through the door she stood up and said, "Now I am well."

She told him the evil spirits had fled, and gave as their reason something the owner of the study found "more kind than true". Mather was so pleased by it that he forbade the experiment being repeated, on the grounds that it was beginning to look like making a charm out of a room.

At family prayers she whistled, sang and yelled to drown his voice, struck at him with her fist and tried to kick him. Her hand and foot always stopped an inch or two short of his body, which was taken as proof that he was protected by an invisible coat of mail. It was proof of nothing except that she knew precisely how far she could go.

When he began writing a sermon about her, she did everything she could to stop him. She knocked on the study door and told him somebody downstairs would be glad to see him; he put down his pen and went down and found nobody. When he scolded her for lying she asked him whether Mrs Mather was not always glad to see him. She threw large books at his head.

He finished the sermon, preached it, and published it. Richard Baxter wrote a preface to the London edition, declaring that anyone not convinced by the evidence that the child was bewitched "must be a very obdurate Sadducee."

Four years before Salem

This is where the case stops being an anecdote and becomes the origin of something much larger.

The Goodwin affair happened in 1688. The Salem trials happened in 1692. Upham's argument, made in the 1860s and never seriously overturned, is that the earlier case is where the later one came from. Mather's account went through the whole country, in print, with a famous divine's endorsement on the front. It gave every household in Massachusetts a script: these are the symptoms, this is what they mean, this is who does it.

The children in Samuel Parris's household in Salem Village were the right age to have read it, or to have heard it read. Four years later they did the same act — the fits, the choking, the accusations, the names — and this time the courts believed it, and twenty people died.

The groundwork was already there. Months before anything was formally charged in Salem, the deputy-governor of the colony was arresting suspected people. Ministers were preaching the alarm. And in Gloucester, that same season, a whole town spent three weeks under siege from spirits it could see: figures gliding over the fields, hovering around houses, appearing and vanishing at the edge of the woods. The inhabitants moved into the garrison. Sixty men came from Ipswich in military array, and several valiant sallies were made from the walls. A great deal of powder was expended and no blood was shed, corporeal or otherwise. The minister who wrote it all down described the enemy's dress — white waistcoats, blue shirts, white breeches, bushy black hair — and hoped that all rational persons would be satisfied that Gloucester had not been alarmed by real French and Indians, but by the Devil and his agents.

A community that has decided what it is seeing will see it.

The other children

Boston was not unique, and the child was not either. In Newbury, the grandchild of William and Elizabeth Morse produced a performance nearly as remarkable and carried it off, deceiving the entire country except one man, Caleb Powell, who saw through it — and got into serious trouble for saying so.

In Amsterdam in 1560, twenty or thirty boys were seized with fits at once, cast on the ground, and left unconscious afterwards, and while the fits lasted they appeared to bring up quantities of needles, pins, thimbles, scraps of cloth, fragments of pots and kettles, splinters of glass and locks of hair. Crowds came to watch. Nobody doubted the Devil was in the room.

The lesson repeats in every one of these cases. The performance is not the difficult part. The difficult part is being the one person in the room who says it is a performance, because that is the person who gets punished for it.

What it actually teaches

Upham's verdict on the Goodwin affair is blunt: Mather was grossly deceived, and audaciously imposed upon by the most precocious cunning he had ever met. He was not a villain. He was a scholar with a theory, and a child who had worked out what the theory needed and supplied it, and a woman at the bottom of the social order who could be hanged for it.

That is the shape of a witch trial, and it is worth being clear-eyed about, because it is not a shape that belongs only to the seventeenth century. The accused is whoever has least room to answer. The evidence is testimony, given by people who are believed because of who they are. And the proof is agreement: the more people who nod, the truer it gets.

It is also worth saying plainly what the child was. She was thirteen. She had found the one thing in that house that made adults hang on her every word, and once she had started, she could not stop — because stopping meant standing up in front of a minister, a magistrates' court and the whole of Boston, and admitting the invisible noose had been a trick. There was no version of the truth available to her that did not end with her disgraced and a woman still condemned.

She was not possessed. She was a child in a room full of adults who badly wanted a demon, and she gave them one good enough to hang somebody.

Drawn from Charles Wentworth Upham, "Salem Witchcraft, Volumes I and II" (1867), which quotes Cotton Mather's own account of the Goodwin case at length. The book is in the public domain and free to read at Project Gutenberg.