Witchcraft did not disappear in Europe when courts stopped prosecuting ‘witches’. In France, this decriminalization was relatively early (1682), and confirmed by the revolutionary law code of 1791. Yet, as historians have long known, the French courts continued to deal with conflicts over witchcraft long after decriminalization. Drawing on digitized newspapers and regional archives, I have built a database of more than a thousand cases touching on witchcraft in France between 1790-1940. This paper concentrates on a subset of 263 cases where suspected witches were attacked or threatened. In 70 of these cases, the ‘witch’ died of their injuries. The database reveals patterns in this violence that neither contemporaries nor historians have noticed before, from the prosopography of perpetrators and victims to the locations where violence was performed. Above all, there is a notable shift from elaborate rituals of torture and burning in the early nineteenth century, to dispassionate shootings after 1850.
Beyond the specific findings about violence and witchcraft in this period, the paper offers some reflections on the broader issues of using of ethnographic or folklore source materials to compare long time periods and different regions, such as labelling. Patterns of violence against witches are not catalogued in newspaper indices or archival finding aids. Historians have often made good use of the categories of the state to chart patterns in violent crimes, but the cultural patterns of witchcraft conflicts are occulted in official categories. Newspaper accounts which did dwell on supernatural beliefs frequently misrepresented them. Some regions have better news coverage than others. Nor did contemporary folklorists have the familiarity – or the desire to gain familiarity – with the repertoires of popular violence to describe these accurately in their writings. The work of assembling and cataloguing cultural examples – as in my database – is always partial. Often, we do not know what we do not know, which poses significant problems to broad comparative approaches.
Close historical work with these primary sources induces a healthy scepticism about the relationship between any such records and wider practices. Some prosecutors and judicial officials tried to cover up the role of supernatural beliefs in criminal cases. Others zealously pursued crimes of ‘superstition’. Trial dossiers are filled with information about the active silencing of the historical record when it comes to shameful or divisive topics. Ambivalence reigns: few participants in violence against ‘witches’ could articulate why they did what they did. Many denied believing in ‘witches’ even after having killed the objects of their suspicion. Trial and news sources are filled with such silences, misdirection, and innuendo, because the meanings of violence were beyond linguistic expression.
The cultural patterns we discern in materials like this are constructions we impose onto imperfect source materials. How much do broad comparisons tell us about the past, and how much do they simply reveal our own research processes and the techniques of archiving and cataloguing they depend upon?