In 1976, in a book on late medieval witch trials, I argued for a distinction between popular and élite or learned conceptions of witchcraft. The documents most likely to reveal popular notions of witchcraft differ markedly, I proposed, from those most susceptible to the bias of judges familiar with current demonological theories. In a significant subset of trial records the original testimony of the accusers is preserved, and these documents refer only to sorcery or maleficent magic, not to the diabolism of the witches’ Sabbath and related notions. Furthermore, the charge of diabolism arises far more often in ecclesiastical than in secular courts, and more often when the records are in Latin than when they are in the vernacular. All of this I took as evidence that sorcery was a datum of popular belief, while diabolism was introduced by learned élites, particularly inquisitors. Accusers and other witnesses believed in sorceresses who bewitched children and cattle, compelled men to love them and rendered people lame. Inquisitors and others who knew the current theological notions about witchcraft and the devil believed in witches who flew through the air to the Sabbath, entered into pacts with the devil, had sexual intercourse with him, and performed maleficia at his behest. If ordinary people did not even believe in diabolism, it is highly unlikely that it was actually practiced.

After more than thirty years, I remain persuaded of the general validity of this argument, but I would phrase it a bit differently and would admit some qualifications. Not long after that book emerged, it became unfashionable thanks to Peter Brown and others to speak in terms of a two-tier culture, to distinguish between popular and élite cultural levels. Yet it still seems to me that when we attend to accusations from the neighbors of the accused we are hearing one set of voices, while the confessions extracted under judicial coercion (most obviously torture) come from a different set of voices. Even in a confession, the voices we are actually hearing are those of the theologian and the inquisitor.

That does not necessarily mean, however, that the accusers and others in society did not believe in the Sabbath. If there are distinct perspectives voiced in the trial records, they may differ less in belief than in expectation.