Elia Nathan Bravo did not believe in witches, not in the classical European sense of  a “sorceress with the power to cast curses thanks to a fidelity pact with the devil.” (Nathan Bravo 2002: 122)  Even more, she was certain that there were no witches, at least as certain as we are that there are no unicorns, and for similar reasons, i.e., we do not have any empirical evidence of their existence. Within the current ontological classification, she would have to be considered a fictionalist, i.e., she held that we ought to interpret the use of “witch”, “bruja” and similar expressions, as used in Europe after the XVI Century, at face value as aiming to talk about, well, witches, that is,  sorceress who derive (at least some of) their supernatural powers from a pact with the devil. This makes most of those statements literally false, since the relevant concept of witch they employ has an empty extension. Thus, she faced the same main challenge that all factionalisms face, i.e., make sense, not of what people say when they talk about witches, but of what they do. Consequently, we can read her work, not just as a study of witch persecution in Europe, but as tackling one of the fundamental questions in ontology – not just social ontology, but ontology in general –: what is the use of empty concepts. 

One of Nathan Bravo’s main claims is that empty concepts are not inapplicable. The distinction between a concept’s extension and its target is useful here. The main moral of the distinction is that the characteristics that makes an object be an X are seldom the characteristics we directly appeal to when trying to determine whether to consider an object as an X. There is usually a divergence, because the metaphysically relevant characters are seldom perceptually perspicuous – a corollary of the widespread truism that nature likes to hide, as Heraclitus reminded us hundreds of years ago. When the divergence is too large, we usually talk of dual concepts in the sense of Prasada, Knobe and their collaborators. But even outside of dual concepts, we usually appeal, not directly to the metaphysically relevant characters that make an object belong to the extension of our concepts, but instead to a set of more accessible characteristics that we take to defeasibly indicate that the object belong in such extension. For example, even people who believe that what makes a dog a dog is its place in a large evolutionary process and maybe some properties of its genetic makeup as well, identify dogs by their external appearance.  I believe that Milú and Tomás are dogs, not because I had a genetic test made or because I know their biological genealogy going back millennia, but because of how they look, sound and smell (oh, how Milú smells – or should I say stinks!) I know that it is not that they are dogs because they are hairy, bark, walk on four legs, etc. yet I identify them as dogs because of these external facts. And we do not need to believe in evil demons to know that these identifying characteristics can mislead us in confusing dogs with other close species.

Now, what happens with concepts like ‘witch’? Nathan Bravo holds that just as they have empty extensions, they also have trivial targets, i.e., at least in principle, almost anyone could have been the target of an accusation of witchcraft. This is so because, just as there is none who has had all the characteristics that define a witch, i.e.,  none has been a sorcerer who also cast spells and also is in a pact with the devil and also derives her powers from such a pact, etc. so almost everyone has one of those characteristics or one that could serve as defeasible evidence for possessing it, i.e., sorcery was common practice, everyday events could easily be interpreted as the effect of curses, etc. Accussations of witchcraft, in Nathan Bravo’s diagnosis, were irrefutable. Thus, no semantic or epistemic approach could not help explain what people did with the concept of witchcraft. It has neither to do with its extension, which is empty, nor with its target, which was trivial and irrefutable. Instead, Nathan Bravo advises us to abandon the individualistic assumption behind such approaches. The relevant causes are not the epistemic states of the individuals involved, neither the beliefs, desires or fears of the relevant clerical and political authorities nor those of the everyday people behind those horrific events, but irreducibly social forces. What Nathan Bravo proposes is a change of focus, from the individual to the social, form the semantic and epistemic to the political and historical. And, as I have also suggested elsewhere (Barceló 2020), from the existential-performative to the social-constructive.