Digital witch-hunts: what online abuse teaches us about the politics of definition

From seventeenth-century trials to twenty-first-century deepfakes, the struggle over who gets to define harm hasn’t changed.

How we define things matters: a definition not only provides us with a shared frame of understanding, but also allows us to properly tell us what something is not. We can then communicate effectively, knowing we all broadly understand what we mean by a fruit or vegetable (OK, we like to debate tomatoes), and people who need to can also conduct research with a shared understanding of terms. For example, better understanding the characteristics of vegetables means that agriculturalists can identify the best conditions for them to be grown (which would be more difficult if we were throwing a bunch of other plant species into the equation).

But definitions can also be slippery. There are many areas in life where definitions are contested: Is wellness a form of healthcare or a consumer identity? Is an ‘expert’ someone who is credentialled through qualifications, or is it simply a matter of being experienced or even being widely followed online? At first glance, the definitions we live by look solid and stable, but when we start probing, they can quickly break down into competing perspectives.

The reason for looking at this topic is to attempt to find ways to understand the way that women, particularly those in public life, find themselves targeted with deepfakes, smear campaigns, and torrents of online abuse. Some commentators have found parallels with the witch hunts of the fifteenth and seventeenth centuries, where thousands of people, overwhelmingly women, were executed as witches across Europe and its colonies. Is it justifiable to define the toxic online behaviours that women face today as witch hunts?

And this is not the only definition question because there is a very live debate around whether current definitions around mis and disinformation properly tackle the torrent of abuse that women online face. Media researchers Marília Gehrke and Eedan R. Amit-Danhi argue that gendered disinformation “has scarcely been addressed by scholars”.

So let’s look at definitions in the context of witchhunts – and examine the basis on which the online abuse women face today might be considered to be in the same definitional grouping as the witch hunts of the fifteenth and sixteenth centuries. And the basis on which this might or might not be considered misinformation.

We explore the implications of this for approaches needed to tackle the alarming issue, as well as what it reveals about the current state of the debate on misinformation.

The witch hunt industrial complex

Between the fifteenth and seventeenth centuries, roughly 60,000 people, overwhelmingly women, were executed as witches across Europe and its colonies. The popular narrative is that this was a case of collective hysteria, driven by irrational crowds. Arthur Miller explicitly framed the Salem witch trials in this way, using them as an allegory for the McCarthyism at the time of writing.

However, it is increasingly understood that these mass killings were not simply acts of mass hysteria, as there is a great deal of evidence that they were, in fact, systematic processes with a high degree of organisation and structure. Theological guidebooks such as the Malleus Maleficarum set out rules for what could be used as evidence of witchcraft. This included testimony that a ‘witch’s spirit’ appeared to the victim, which was to be admitted as proof, and that any denial itself was read as guilt. Another set of rules involved interpreting misfortune as the result of witchcraft, so a failed harvest became proof of ‘cursing’. These guidebooks offered a receptive audience a set of mechanics that allowed rumours to readily be converted into guilt and, from there, into punishment.

The resulting trials were not only legal procedures but public spectacles designed to make persecution of witches visible to a wide audience. Pamphlets carried sensational accounts of witches’ confessions, which allowed suspicion to spread beyond the courtroom. Public executions provided spectacle, with bodies destroyed before crowds to dramatise the triumph of authority over evil. Witch trials served as public lessons in godly behaviour and social discipline, in a sense operating as platforms for authorities to communicate moral boundaries to the wider community.

And historians point to the way this machinery was useful for those in positions of power. For vicars and priests, witchcraft claims drew a scattered set of lay beliefs into a narrative about evil with the church at the centre of interpretation. For magistrates, it expanded their reach, as manuals and “experts” ensured that rumours and neighbourly disputes were translated into evidence that was admissible in court, allowing the state to claim authority over what would otherwise be unseen. For rulers, it offered political expedience: crop failures, epidemics, and fiscal crises could be attributed to hidden enemies, thereby redirecting resentment away from institutional failings.

Of course, as political philosopher Silvia Federici argues, misogyny was at the heart of this: the social ordering of the period made women disproportionately vulnerable to accusation. Midwives and healers held reproductive knowledge that challenged the emerging professional power of male physicians, and women who deviated from prescribed norms of obedience, chastity, or economic dependence were recast as dangerous. In Federici’s famous book, Caliban and the Witch, the figure of the witch serves as a vehicle for disciplining female autonomy and reshaping the moral and economic order.

While these may seem part of history, accusations and witch-killings still occur in parts of sub-Saharan Africa, South Asia, and Melanesia, with victims again predominantly women.

Digital hunts: deepfakes and the manosphere

Returning to our question of definitions, has the witch hunt now taken a different form, one that operates through digital media?

In some ways, it is helpful for campaigners to point out the parallels, as the mass slaughter of women at that time is something for which there has been little acknowledgement by those in political power (with the notable exception of the Scottish Government). But also the extent to which women are subject to violence by men is huge: an estimated 736 million women (almost one in three) have been subjected to physical and/or sexual intimate partner violence, non-partner sexual violence, or both at least once in their life (30 per cent of women aged 15 and older).

And there is a huge amount of non-consensual deepfakes and other forms of image-based sexual abuse (IBSA) overwhelmingly targeting women with devastating impacts. The parallels are easy to see: just as early witch hunts were staged for maximum visibility, deepfakes and doctored images are designed for mass display. The resulting humiliation is a spectacle, implicitly teaching audiences to read women for evidence of wrongdoing or shame.

But there might also be risks in defining digital misogynistic abuse as a “witch-hunt.” Early modern trials were legal and religious events, authorised by church and state, whereas today’s digital persecutions are driven by platforms, anonymity, algorithms, and networked misogyny. The analogy is useful for highlighting how gendered accusations, spectacle, and public shaming persist, but treating them as a single category perhaps risks pointing us toward the wrong solutions. As writer Sara Ahmed flags, moral recognition without structural repair can provide the illusion of justice.

Structural orchestration: the social and systemic forces across the centuries

There do seem to be parallels between early modern witch-hunts and digital misogynistic abuse, both involving patterns of accusation, spectacle, and gendered punishment. On the other hand, it could also be argued that today’s online misogynistic abuse is not a witch-hunt in the strict sense: it lacks the codified rules and formal pathways through which historical hunts converted accusation into punishment. On this basis, it could be considered to sit closer to a smear, an attack that turns ambiguity into stigma.

But definitions are not fixed; they are choices about where to look and what to make visible. A lens that focuses on the mechanics of these accusations draws on a cognitive or information-processing perspective, concerned with how claims are formed, spread, and judged. By contrast, a lens that centres the victim (rather than the process) shows more clearly how women are positioned as dangerous or deceitful, making them ready targets. This broader socio-cultural lens means we can trace the narratives and power relations that determine the way accusations form and spread in the first place, a more fundamental understanding (rather than the mechanics of the process).

To examine witch-hunts using this socio-cultural lens, we can turn to writer Sian Norris, who argues that misogyny is anchored in a zero-sum belief: women’s gains are men’s losses. In the context of deindustrialisation, the rise of precarious labour, and the retrenchment of welfare, this belief reframes laws on workplace equality or protections against gender-based violence as attacks on men. International polling underscores the point.

In an Ipsos survey of 22,000 adults across 32 countries, more than half agreed both that women’s rights have gone “far enough” and that men are being asked “too much” to support equality: a claim that is clearly at odds with persistent gaps in pay, safety, and political representation.

Here we can see more clearly parallels in the patterns at work. In the past, crop failures and disease were blamed on women who were already viewed with suspicion; today, the data suggests that feelings of status loss and economic insecurity are often redirected toward feminism or women’s independence.

In both cases, misogyny provides the lens through which problems are interpreted, and it is then that the surrounding infrastructure helps those accusations spread. Of course, these are quite different, thus masking the underlying similarities: historically, it was manuals, courts, and local authorities; today, it is algorithms, platforms, and online networks.

Situating witch-hunts within a wider context of misogyny, not just as a type of accusation regime but as part of a longer cultural pattern in which women are cast as the source of others’ losses, we can see why definitions are not a side issue but the substance of the problem.

Where does misinformation sit?

So why does it matter if we find a parallel between the witch-hunts of old and the digital abuse of today? The answer is that if we can name these continuities, then we make them politically visible, allowing us to see how the same logics of suspicion, shame, and silencing are still at work.

Calling today’s digital abuse a witch-hunt gives it moral weight, aligning it with what is widely recognised as one of history’s most shameful chapters (though in some parts of the world, it is not yet history at all). The point is not only analytical but political: naming creates resonance. It reframes what might otherwise be dismissed as private cruelty or online hostility as something collective and structural, an enduring pattern of control.

On the same basis, it is worth considering that witch-hunts can sit within the broader category of misinformation, as this would offer visibility and policy leverage (given the policy focus of this field). However, we quickly encountered a problem: if we define misinformation through a process lens (the dominant frame used to define misinformation), which focuses on how information is produced, transmitted, and consumed, then today’s gendered digital abuse does not easily fit. Within an information-processing framework, misinformation is typically understood as false or misleading content that distorts belief or knowledge. The focus lies on the mechanics of diffusion and cognition: how people encounter, assess, and share claims.

From this vantage point, misogynistic attacks online appear less as “misinformation” and more as smears, as their harms arise through narrative framing and amplification rather than factual inaccuracy. As Deen Freelon and Chris Wells observe, mainstream approaches to misinformation often privilege truth over consequence, treating it as a failure of knowledge rather than an expression of power.

Seen through that lens, the online vilification of women falls outside the definitional frame. It manipulates stereotypes and emotion, rather than verifiable facts. Yet, as Rachel Kuo and Alice Marwick argue, this narrow reading of the issue reflects a depoliticisation: focusing on accuracy, misinformation research often ignores the power structures that give false claims their force.

And this is why definitions matter. A definition of misinformation that is based on cognition and falsity means that its gendered dimensions disappear, as the problem becomes one of individual reasoning. A socio-cultural lens can restore what that framework omits: the histories, power relations, and moral narratives that determine whose voices are trusted and whose truths are dismissed. When online abuse of women is excluded from misinformation frameworks, the result is a blind spot: it contributes to an ‘invisibalisation of these situations of abuse that require specific and targeted responses.

Conclusions

What this analysis shows is that early modern witch-hunts and today’s digital targeting of women can appear quite different if we focus only on the surface: the processes, the institutions involved, through a largely cognitive or procedural lens. But when we take a socio-cultural perspective and view them from the standpoint of the person being accused, the alignment becomes clear. A socio-cultural lens brings this into focus: it exposes the misogynistic narratives, patriarchal expectations, and long-standing credibility gaps that mark women as inherently suspect and therefore as believable targets across time.

A cognitive lens is still needed. It helps explain how people react to accusations, how rumours spread, and how uncertainty shapes judgement. But while it seems necessary, we have made a case, it is not sufficient. On its own, it focuses attention on individual reasoning and risks obscuring the cultural and historical forces that decide who is targeted and why. Epistemologists such as Miranda Fricker help explain why this gap matters. When women’s experiences of abuse are treated as peripheral to misinformation, it reproduces what Fricker calls epistemic injustice, the systematic devaluing of some people’s voices as credible knowers. In this sense, excluding gendered online abuse from misinformation frameworks is not just an oversight in policy but a continuation of that same credibility imbalance.

This is why we need a plurality of lenses. Cognitive perspectives explain mechanisms of spread; socio-cultural perspectives explain meaning, power, and vulnerability. Taken together, they demonstrate that the continuity between past witch-hunts and contemporary gendered attacks does not lie in identical procedures but in enduring societal structures of suspicion, misogyny, and harm.

The question, then, is not simply whether today’s attacks “count” as witch-hunts. It is whether our definitions help us see the pattern clearly or whether they leave women’s experiences out of the frame. With narrow definitions, the danger is that we look in the wrong places. With plural, victim-centred ones, the pattern becomes much clearer and far harder to dismiss.