Stochastic terrorism: critical reflections on an emerging concept

ABSTRACT

This paper critically discusses the recent concept, stochastic terrorism – broadly, the idea that influential individuals may demonise target groups or individuals, inspiring unknown actors to take up terroristic violence against them. I collect together different strains of thought on the emerging concept, reflecting critically on what a suitable definition of the phenomenon would look like (or whether it would be needed), what the social urge to coin the concept may reveal about authoritarian power and violence, and what makes this form of political violence possible. I argue that present commentary fails to emphasise sufficiently the role of mistruth and deceit in such rhetoric, as well as its historical and mainstream precedents. Moreover, I understand the phenomenon to be specifically authoritarian in nature, which not only demonises but dehumanises its targets. In light of this, I suggest that given both the mainstreaming of racist conspiracy theory and the historical and continuing presence of centrally constructed “folk devils”, the authoritarian problem which can manifest into stochastic violence is very much endemic to modern liberal democracies. With this framing of stochastic violence in mind, we ignore it simply as a buzzword at our peril: even if the theoretical issues I have highlighted continue to persist, even if we struggle to pin down the concept with desired clarity – it bears a phenomenological significance and reflects an ongoing political structure of violence.

KEYWORDS:

• Right-wing terrorism
• far-right terrorism
• stochastic terrorism
• critical Terrorism Studies
• terrorism studies

Introduction

An emerging concept recently “bandied about” (Amman and Meloy 2021, 2) in discourse about security and counterterrorism, describing a “a new form of political violence” (Braddock 2020, 210), is stochastic terrorism. The concept has been mentioned or used in criminological (Hamm and Spaaij 2017) and anthropological (Biondi and Curtis 2018) literature, policy documents (Heine and Magazzini 2020) and counter-terrorist work (Allchorn 2022), and by former national security personnel (Kayyem 2019, 2022). But the discussions which employ this concept do not always define or characterise it, and when they do so, not all characterisations are together consistent. Considering that scholars in terrorism and security studies (orthodox and critical), criminology, sociology, and beyond might want to investigate it further, and that civil society or citizens may have concerns about the threat which the concept purports to detail, clarification is needed on the shape and usefulness of the concept, “stochastic terrorism”.¹

Amid the concept’s limited academic treatment thus far, the criminologists Mark Hamm and Ramón Spaaij dedicate a section to stochastic terrorism on their work on “lone wolf” terrorism, where they describe stochastic terrorism as “the use of mass media to provoke random acts of ideologically motivated violence that are statistically predictable but individually unpredictable” (2017, 84). In discourse that focuses on North America, Europe, and Australasia, the concept continues to gain traction, given perceived or actually increasing security threats from the far right. In particular, the words of former President Donald Trump on the precipice of the “Storm on the Capitol” on January 6th, 2021 are sometimes held up as an exemplar (e.g. by Amman and Meloy (2021)) of stochastic terrorism: an influential speaker – who becomes the stochastic terrorist – on a podium uses language that may be taken (or was taken) by an audience to encourage them – the perpetrators – to political violence. Accordingly, this notion as standardly discussed holds that there are two categories of terrorist involved: first the stochastic terrorist, then the perpetrators or “lone wolves”. But it remains unclear how best to, or whether even we ought to, characterise this notion, and what we would be committed to in so doing.

Beyond a simply clarificatory mission, in this article I suggest a qualified embrace of the term – or ultimately, perhaps, a related term (“stochastic violence”) – arguing that on analysis it may serve as a useful and important directive to examine authoritarian structures within democratic societies past and present. In the opening section, I overview the concept and example cases, critically enquiring about the unclear role of intent in present characterisations. Next, I argue that any reformulation of the idea must reference normatively substantive elements – i.e. whether the “stochastic” speech is true, false, misleading, and so on. Applying this thought in the second section, I recommend a strong focus on dehumanising rhetoric, arguing that a larger array of background factors that make this violence possible then comes into view. These include conspiracy theories, whose potential to demonise subjects I explore as grounded on the mainstream vilification of “folk devils” in mass media and by government campaign. Lastly, I reflect in the final section on what we must keep in mind by using a notion of stochastic terrorism, especially as raised by critical terrorism research more broadly. In particular, I note the risk that Jackson’s “epistemological crisis” persists within this newer notion, but remain undeterred about its potential for useful analysis.

The concept of stochastic terrorism

The concept of stochastic terrorism is usually traced back to a 2002 piece by risk analyst Gordon Woo (e.g. in Braddock (2020) and Amman and Meloy (2021)), whose use of the term is cursory and not clearly defined. The turning point in the concept’s usage is a weblog post authored by a person known only by their pseudonym, G2geek (2011). The tone of this piece, perhaps befitting a blogpost, is somewhat polemical, but the content has largely set the contours for this concept ever since:

Stochastic terrorism is the use of mass communications to stir up random lone wolves to carry out violent or terrorist acts that are statistically predictable but individually unpredictable (G2geek 2011 – emphases original)

The post goes on to clarify, in a usage that has now solidified, the perpetrator is not the “stochastic terrorist” – that, rather, is the communicator who inspires the perpetrator to act. In expanding on the definition, the author draws parallels in the use of this method between Osama bin Laden and particular Fox News television anchors on the US right wing: Glenn Beck, Bill O’Reilly, and Sean Hannity. The “stochastic” element is a notion borrowed from mathematical statistics; the italicised portion of the quote above is, more or less, what it is taken to mean in uses of the phrase.

It will help to have a concrete example. In their book on lone wolf terrorism, Hamm and Spaaij (2017, 84–90) outline some of the same examples of stochastic terrorism as the blogpost, including the case of Byron Williams. In 2010, Williams, in San Francisco, had been on his way to committing mass homicide at the Tides Foundation and the American Civil Liberties Union when apprehended after a shootout with highway police forces. According to Hamm and Spaaij (2017, 85–86), those close to Williams described his avid watching of Fox News, and especially Glenn Beck’s show; Williams cited Beck specifically as his “primary source” for some of his beliefs concerning the alleged nefarious activities of these organisations. (Beck had verbally attacked the Tides Foundation 29 times in the 18 months prior to Williams’s attempted mass homicide.)

At a conceptual level, there is much to flag and discuss about the idea of stochastic terrorism as so far expressed. In this section, I focus on two elements to begin clearing the ground: the unclarity of speaker intent and the normativity of their claims.

Intent

At first glance, the basic supposed structure of stochastic terrorism applies easily enough to the Williams case: a speaker, Beck (the stochastic terrorist), uses a mass media platform (Fox News) to broadcast incendiary rhetoric over time, seemingly bringing an audience member, Williams (the perpetrator/lone actor), to the point of wanting to commit a violent, politically inspired attack against a particular target (Tides Foundation/ACLU). Moreover, it is plausible to think that these persistent verbal attacks over time would produce this sort of outcome (or attempted outcome), without knowing precisely who would ultimately construct and enact a violent plan. So much, so simple.

But the concept as outlined should give us pause. It is rather unclear what it takes to rightly hold a person or group as providing this “stochastic” inspiration, and thus to invoke questions of intent (and, later, of legal or moral responsibility). The blogpost author had been straightforward: intent enters the picture here because the attacks predictably follow the rhetorical outbursts of these figures, and because the corporate media network concerned understands, the author claims, this relationship between rhetoric and violent outcome. Hence, because “it takes something between callous disregard and deliberate intent” to continue the rhetorical programming, Fox News or Glenn Beck, in G2geek’s examples, has both intent and moral responsibility. But it is less clear where Hamm and Spaaij stand. Consider their next case study of stochastic terrorism: in 2012, Floyd Corkins, a gay rights activist in the US, unsuccessfully sought to commit mass homicide at the Family Research Council (FRC). Corkins chose the FRC as his target because the Southern Poverty Law Center (SPLC) had listed them alongside other anti-gay groups: in 2010, they called the FRC a “font of anti-gay propaganda throughout history”, locating its headquarters on a “Hate Map” (as cited in Hamm and Spaaij (2017, 87)). Subsequent to the attack, the target (the FRC) called the SPLC “reckless” in giving organisations the label of hate groups. Hamm and Spaaij then neither signal agreement nor disagreement with this claim; perhaps it stands there as an allegation of a kind of reckless intent – how else does this case qualify as stochastic terrorism? Absent any more concrete claim, it remains unclear what the role of intent here is supposed to be, and just how fine-grained it might have to be to qualify an actor as a stochastic terrorist. Elsewhere, Molly Amman and J. Reid Meloy, writing on stochastic terrorism, seem fairly relaxed about the matter: “The intent of the speaker to cause such violence may range from unwitting naivete – in the sense of an accidental sin – to full knowledge and hope that such violence will happen, the risk magnified by his or her public speech” (2021, 4 – emphases added).

Within critical terrorism studies, although there is a struggle over how to characterise intent in terrorism as a more general concept, it is normally present in some capacity. For example, the appendix to Alex Schmid’s article on the “definitional problem” (2004, 418–19) lists 11 definitions, and some notion of intent crops up in over half (while arguably implied in the others). Zulaika likewise agrees, in passing (while accepting nonetheless that intention must always be further interpreted): “intentionality is in itself a criterion for terrorism” (2016, 41). Consider also Richard Jackson’s discussion of a “soft” notion of intent, which “assumes that, if actors undertake actions for which they can be reasonably assured of the outcomes, they therefore intended those outcomes and can be held responsible for them” (2011, 124). Jackson suggests in practice it is then not difficult to infer intent in individual cases, as we cannot seriously be credulous in the face of an agent disavowing his intent, knowing he has, say, directly attacked a crowd. Of course, such an inference is supposed to be relatively harder in the case of stochastic terrorism, something to which the prominence of “plausible deniability” in characterisations of it attests (e.g. see Braddock (2020, 224)) – the stochastic terrorist is thought of as a person on whom we might, as observers or commenters, struggle to pin the violent intent, perhaps given the use of “dogwhistle” language, or just by virtue of their standing further back, so to speak, from the acts of violence themselves.

One consideration in favour of including at least this soft intent concerns inflation: without intent, might we otherwise end up classifying far too many acts of violence as stochastic terrorism? More concerningly still, perhaps it would become easier through this rhetorical device to classify as terroristic the words and speeches of power’s traditional political targets: ethnic minorities, academics, journalists, LGBTQIA+ communities, and more. Indeed, an instructive case here arises from Hamm and Spaaij’s example. Leaving intent out the picture (as they do), should the Floyd Corkins incident count as an instance of stochastic terrorism even if the SPLC was correct to call one of Corkins’s targets a hate group? If so, we now tread in murky waters, for it seems as though parties are at risk of committing indirect terrorism simply for labelling any institution, organisation, or group as hateful, given that others may act violently on that characterisation. The paradox this leads to is that we may no longer designate the groups most consistently demonstrating hateful attitudes towards others as hateful, for, by doing so, we could “become” terrorists (when the hate group is or could be targeted as a victim in turn). In this connection, it is especially important not to forget that the “terrorist” label has moral and political repercussions; as Jackson notes, for those so labelled, it “has real consequences for their lives and well-being, as well as that of the community from which they emerge” (2011, 119; see also Sageman 2017, 12).

Hamm and Spaaij are illustratively incautious in how they discuss this case. They note that gay rights activists had called for boycotts of and protests against another of Corkins’s targets, the chain restaurant “Chic-fil-A”, given the restaurant CEO’s open opposition to same-sex marriage (2017, 87). Corkins had wanted to smear sandwiches from this restaurant on his would-be victims after killing them, an act Hamm and Spaaij summarise as “apiece with this gay rights protest,” adding that “the fifteen Chic-fil-A sandwiches were symbolic of his commitment to the cause” (2017, 87). In describing Corkins’s planned attack as “apiece” with the civil rights protests, Hamm and Spaaij may implicitly locate the stochastic actor(s) within the protests or as the protest movements themselves. But this now makes terrorists (albeit stochastic ones) of broadly progressive and peaceful protest movements, simply by virtue of their even truthfully identifying the threats that other groups pose to them. Proposing that civil rights protestors who accurately articulate a threat towards a minority or out-group might unwittingly, by articulating, become stochastic terrorists is, at the minimum, unsavoury and counterintuitive. Whatever such indirect encouragement could amount to, it surely could not sit on all fours with examples that involve the hateful rhetorical targeting of subjects to an audience known to comprise among its number very unstable and easily influenced individuals.

Normativity

Notwithstanding that clarity on the role of intent in existing discussions of stochastic terrorism would be desirable, there is another route to sharpening the concept to prevent its becoming too broad or so easily abused. What goes awry in Hamm and Spaaij’s analysis in the Corkins case is not so much their lack of focus on intent, but that the normative substance of a text or speech (in the sense of descriptive, not moral, rightness and wrongness) is nowhere in view. We lead ourselves astray if we miss the relevance of this to the phenomenon. That is, the characterisation of stochastic terrorism – especially if it does away with intent altogether – should include whether some candidate speech contains true, false, or misleading statements (or uses linguistic techniques to imply, implicate, such thoughts, etc.). A similar thought may lie behind why other definitions often mention demonisation, such as that by Dictionary.com, who (after seeing a 63,389% rise in searches for the term following a mass shooting in Texas in 2019) defined “stochastic terrorism” in a special feature as “the public demonization of a person or group resulting in the incitement of a violent act, which is statistically probable but whose specifics cannot be predicted”.² Their primary source here is Juliette Kayyem, a national security expert formerly of the Obama administration’s Department of Homeland Security; Kayyem elsewhere defines stochastic terrorism as “the demonization of groups through mass media and other propaganda that can result in a violent act because listeners interpret it as promoting targeted violence” (Kayyem 2019).

But sticking to my thought that normatively substantive content is pivotal, we can see that even “demonisation” is not sufficiently fine-tuned, given that its definition tends to be normatively neutral: to demonise is simply to portray a group or person as a grave threat, not to rightly or wrongly do so. Consider here two examples germane to the worry. Marine Le Pen’s rebranding of her party, to rid it of (correctly derived) associations with Nazi sympathies, has been described by both media and her party as “de-demonisation”, from the French dédiabolisation (Mondon and Winter 2020b, 33; Trippenbach 2022). The Le Pen rebranding reflects a broader tendency as seen in the victimisation narratives of the so-called alt-right, such as when Paul Joseph Watson (a figure recently exposed as holding deeply anti-Semitic views) accused mainstream media of “anti-white” institutional racism: they “stigmatise and demonise an entire group of people for the actions of a few individuals” – specifically white people – “I refuse to be demonised for the colour of my skin”, as cited in Byline Times (2022 – emphases added). (See Renton (2021, chap. 11) for more on far-right victimisation narratives, especially in the context of speech and its silencing, and Allchorn (2022, 40–41).)

So an adequate characterisation of stochastic terrorism ought to include more than just normatively neutral demonisation. The ground of the appeal to demonising is surely to capture those maligning cases; one of the harms of demonising – to portray, say, a minority ethnic group as a grave threat to the majority ethnic population – is doing so by exaggeratory, propagandistic, or deceptive means. Without suitably narrowing our working definition of stochastic terrorism to include only those cases of wrongful and deceitful demonisation, too much would fall under its umbrella, such as the gay-rights protest movement and arguably the SPLC in the Corkins case above. Consider: conscientious objectors and whistle-blowers reporting the war crimes of powerful regimes; journalists and scientists reporting the threat which organisations who perpetuate climate breakdown pose to a safely inhabitable earth; civil rights activists protesting against violent and wrongful discrimination – all could conceivably be said to demonise an organisation or individual to some extent. Moreover, a person who commits, or aims to commit, violence against a demonised individual or institution may often be inspired by such demonisation. But this fact ought not in and of itself lead us to characterise the demonising party as terroristic. Accordingly, any definition of stochastic terrorism which depicts the relevant relationships in a normatively neutral fashion is liable to provide limited utility for understanding the phenomenon, since too many cases of ideologically inspired violent acts would meet the definition. In doing so, it would encourage authoritative overreach for states and powerful individuals – for it is not just political demagogues with platforms, but academics, journalists, civil society actors, and activists whose influential views might inspire (or be portrayed as inspiring) a person to political violence.

We ought, then, to characterise the concept of stochastic terrorism in a normatively substantive manner by making essential reference to the truth or falsity (or else the misleadingness or deceitfulness) of the putatively stochastic speech. Insofar as we cannot interrogate or apply “stochastic terrorism” without sharpening its edges, here is a defeasible characterisation: we may classify an act of political violence as stochastic terrorism only if the relevant inspiration for the attack involves the false or misleading demonisation of the attack’s target. (I have used the active voice and a subject here as a conscious nod to Sageman’s (2017, 9ff.) point that “terrorism” picks out a categorising process that groups of people perform.) Beyond being normatively substantive, this necessary condition likely becomes politically substantive in turn, on account of the tendency of particular political regimes and actors precisely to play fast and loose with the truth. (For example, see Stanley’s (2018, chap. 1) characterisation of fascism as involving a displacing of truth with myth.) Of course, that includes authoritarian political formations of different political persuasions, as well as some formations of radical populism.

Now, focusing on truth and related concepts here invites a range of critical questions. What is true; who decides; what about those mistakenly speaking falsehoods; what about competing accounts of the facts, or even conceptions of the truth? “[T]he anxiety about talk of truth [in politics] remains great: that once any such talk is allowed through the door, it must bring with it a history of metaphysical baggage and a future of political domination”, as Jeremy Elkins and Andrew Norris put it (2012, 2). The question over arbiters of truth in the context of political speech is massive, given so much everyday political speech is exaggerated, omissive, and contentious. Were my account of stochastic terrorism required or intended to filter down to empirical work or policy initiatives, this would become a larger concern. In criticising Cassam’s (2021) normative approach to understanding extremism, Rik Peels expresses such a worry: “we should probably be able to operationalise [Cassam’s] core definitions, [but] exactly how can we do so if they are normative all the way down?” (2022, 1039). But since my project here is chiefly theoretical, I place no great weight on its being easily operationalised (though note Cassam’s reply (2022, 1046–47) that normative concepts can do empirical and political work).

The bigger issue is that failing to note a particular orientation towards/away from truth would simply not be engaging properly with the phenomena as they present. Take two instances of political speech which “demonise” a target, that in some sense have inspired violence among the audience – parking intent, agreeing it is unclear who intended what – a key means of distinguishing these cases, even to the point of assigning responsibility, lies in discerning the normative direction of the demonising. This is something we do and have a practical capacity for, much as we can discern the selling tactics of marketeers and know when we’re being fleeced. Hence my objection to Hamm and Spaaij allowing for the thought that SPLC or gay activism had stochastically inspired Corkins’s subsequent attack. Consider another example: left-wing groups of the 1970s/80s agitating against capitalist oligarchs.³ If the rhetoric of such groups putatively inspired violence but consisted of such true claims as those about, say, capitalist exploitation of labour, differential ownership of the means of production, and so forth, then we could hardly look to the speaker(s) as providing anything inflammatory to violence. On the other hand, were their claims (poised in front of an audience with heightened fear and worry) to include that workers in some factory had been beheaded over minor transgressions, that the bosses were not merely selfish but also animals, virus-like, or a different “race” or species, and so on – in such a scenario the language used is far more obviously inflammatory, and it is so because of the exaggerations and outright falsehoods it contains as part of the demonisation (the sort of falsity in speech, further, that is not uttered mistakenly). Moreover, this element of the phenomenon explains why, as I discuss in the next section, dehumanising is central to stochastic terrorism.

The stochastic violence of the mainstream

Having refined the initial idea of stochastic terrorism to cases where speakers wrongfully demonise targets, we can now spot its connections with other relevant rhetorical patterns. Most obviously, dehumanising rhetoric comes into focus as a vital and clearly wrongful demonising process. And since dehumanising is a feature of authoritarian movements and structures, we can position stochastic terrorism as a tactic of authoritarianism to understand its larger political and violent role. But it would be an error to confine the notion just to outwardly authoritarian systems, since authoritarian ideology structures mainstream institutions and is proliferated through mainstream information channels – notably by the use of “moral panics” about “folk devils” (or demons, we might say). Accordingly, in this section, we focus on the dehumanising impulse that reverberates through a wider structural frame, enabling the violence captured by the notion of stochastic terrorism.

From demonisation to dehumanisation

Prior to the incident at the Capitol Building in Washington, USA, on January 6th, 2021 (see Amman and Meloy (2021) for a discussion of this incident as stochastic terrorism), Donald Trump spoke in front of a large crowd of supporters. While much interest about Trump’s speech inevitably focuses on whether “And we fight. We fight like hell. And if you don’t fight like hell, you’re not going to have a country anymore”⁴ constitutes an incitement to violence, insofar as we are analysing it as a case of stochastic terrorism, it is not the most important or interesting talk. For the speech is littered with negative characterisations of, and false claims made about, those who became the targets of the subsequent mass mob – Democrat and “weak Republican” politicians (and their offices). Consider here such claims as “our election victory stolen by emboldened radical-left Democrats”, “For years, Democrats have gotten away with election fraud and … There’s so many weak Republicans”, and “Make no mistake, this election was stolen from you”. Another significant example from Trump is his simple labelling of the COVID-19 virus the “Chinese virus”, which not only associates the idea of that virus with Chinese people, but also the more general notion of infection; it is a dehumanising tactic. Sadly, “[a]nti-Asian sentiment and behaviours swiftly followed, with assaults on Asians and Asian Americans being reported” (Bloodsworth-Lugo and Lugo-Lugo 2023, 183).

Wrongful demonisation, and especially dehumanisation, of targets is a core strategy of authoritarian politics. Dehumanisation is a representational process which is, intrinsically, normatively defective: humans are human, not subhuman or (merely) animal, and it is surely this more than any other aspect of problematic rhetoric which portends worrying consequences (cf. Braddock (2020, 224), Wilson (2018), and Sageman (2017, 8) for emphasis also on dehumanisation). Indeed, in their “continuum of violence” hypothesis, Nancy Scheper-Hughes and Philippe Bourgois (2004, 20–21) place dehumanisation alongside other, normalising “everyday” forms of violence as factors that enable a “genocidal capacity among otherwise good-enough humans”. Accordingly, it is unsurprising that O’Connor, Lindekilde, and Malthaner emphasise this element in the process of radicalisation, which “evolves when select individuals or groups are vilified and dehumanised, leading violence against them to be not only normalised but also exhorted as a virtue necessary to protect one’s community or people” (O’Connor, Lindekilde, and Malthaner 2023, 214). In this connection, consider how Karina Biondi and Jennifer Curtis position the “lone” actors that commit terrorist violence within larger authoritarian political structures:

Non-state actors and lone wolf terrorists are tactically useful for authoritarian movements, and working in synergy with politicians and political parties, their violent actions extend the repression of authoritarianism beyond the prison and border to the quotidian spaces of synagogues and supermarkets (Biondi and Curtis 2018, 47)

Their analysis treats the modern examples of incendiary rhetoric inspiring domestic terrorist acts as part of a larger trend in US domestic terrorism. Dehumanisation of target groups is key to their concerns about what they term stochastic violence, chiefly in the rhetoric first of candidate stochastic figures such as former US President Donald Trump or former Brazil President Jair Bolsonaro, and second in the online and offline discourse of their followers.

The US-based examples of stochastic violence Biondi and Curtis discuss include, all in 2018, Cesar Sayoc’s mailing of pipe bombs to political figures that Trump had attacked, including Democrat politicians and personal critics in law enforcement and Hollywood (Weiser and Watkins 2019); a chapter of far-right group “Proud Boys” attacking Manhattan citizens around the same time period; a white nationalist gunman, Gregory Bush, murdering elderly black citizens in a supermarket; and Robert Bowers shooting and killing 11 people at a Pittsburgh synagogue. Bowers had posted on social media much anti-Semitic content, including the conspiracy theory that Trump and others in his administration pushed claiming Jewish philanthropist George Soros was funding caravans of migrants south of the US to “invade” across the border (Craig, Berman, and Achenbach 2018). These cases are ones through which we cannot always draw a very straight, direct line between the rhetoric of powerful, well-broadcast speakers and subsequent mass homicides. The crucial connecting tissue is the idea not (or not only) of words that cause actions, thereby problematising chiefly the speech or the intent behind it, but a larger authoritarian structure in which the dehumanising rhetoric finds a home. Thus, as I take it, their use primarily of the phrase stochastic violence rather than terrorism is one that invites us to problematise a larger array of social dynamics and structures than we might at first think.

Despite Biondi and Curtis’s use of the phrase “lone wolf”, relevant here is criticism of that very idea – that it individualises a type of crime or violence that is usually shot through with ideology (see, e.g. Schuurman et al. (2018, 2019) and Bouhana et al. (2018)); moreover, that the nominally lone wolves in question may have comrades in the form of party or group affiliation and given shared ideologies, so that the phrase acts to shift “attention away from the social character of language and political narratives” (Berntzen and Sandberg 2014, 760). In my view, there is a clear risk that the label of stochastic terrorism and its application only indirectly reinforces that fictitious aloneness, even as it seeks to disrupt it by expanding the focus from the perpetrator to the stochastic terrorist. Some uses of the notion may just individualise another actor in turn (we only need to think of Trump fixations here to reinforce the point), obscuring from view just how essential to the crimes in question are a range of structural factors.

Therefore, to make this phenomenon of stochastic violence intelligible, we ought to see it as a tool of authoritarian politics which interacts with pre-existing elements of many liberal democracies: a sensationalist and sometimes bigoted media ecology (which Hall et al. (1978) explore so thoroughly), interacting with a mainstream policy environment that includes political parties as well as media outriders and think tanks;⁵ a citizenry liable to conspiratorial thinking; large social media platforms and specialised online forums and messaging applications; mediated political discourse, including “culture war” debates that seek to position the authoritarian right or its members as victims; and plausibly, also, increasing material inequalities. It is some such (non-exhaustive) network of factors that makes this form of political violence possible.

But it is constantly tempting to let these wider factors slip from view, especially when we apply narrower analytical models. Amman and Meloy’s (2021) work on stochastic terrorism is a case in point. They aim to depict “the cognitive and emotional impact of demagogic rhetoric upon the receiver or listener, and the movement from idea to violent action” (2021, 9). To do so they use the linguistic theory of conversational implicature to analyse the speaker’s language, a psychoanalytical model to explore the vulnerability of the listener(s) in large group dynamics to narcissistic leaders, and terrorist risk assessment tools to outline the character of the criminal actor. By applying these theoretical frameworks, Amman and Meloy make interesting suggestions about the stochastic terrorist and the listeners they inspire to violence. For example, they position stochastic speech plausibly within the broader theory of threat discourse (2021, 3–7) – the ordinary use of political speech to coerce desired outcomes – marking it out as unique for coercing specifically violent solutions to threats.

Work on lone actor radicalisation also nods to a mechanism between speaker and violent outcome, but its discussion edges closer to the wider structural factors which, I argue, make the most sense of stochastic terrorism.⁶ “Encouragement cues”, initially the subject of experimental psychology some decades ago, are “situational or contextual factors which will lead the individual to express his attitudes behaviorally” (Abelson 1972, 26). In this formulation, they seem unhelpfully local and individualised; they prompt individuals to perform certain actions, e.g. by seeing them modelled by others. But work on the radicalisation pathways of lone actors discusses them in connection with the ideas of stigmatisation and legitimisation, thereby invoking something more normative and social. For example, regarding a 2020 far-right attack in Hanau, Germany, in which the lone acting perpetrator fatally shot several people including those at two shisha bars, O’Connor, Lindekilde, and Malthaner point to the party Alternative für Deutschland previously campaigning at both the national and regional level to “identify” shisha bars with criminality, including gang rape: “Such utterances by institutionally legitimate politicians and parties whose comments enjoy massive public reach can serve as indirect encouragement cues for lone actors to engage in violence” (2023, 221). It is unclear to what extent we can rely on this notion of an encouragement cue qua causal mechanism; indeed, the authors note the difficulty of establishing “direct causality” between political speech and political violence. Nonetheless, they express their concern over the trend of “open endorsement and promotion of views that identify and stigmatise specific communities as legitimate political targets” (O’Connor, Lindekilde, and Malthaner 2023, 221), which rightly speaks to the social significance of an encouragement cue.

Empirical findings about lone actor radicalisation are further instructive in taking us away from individualising tendencies, given the lack of singular, replicable pathways and profiles (O’Connor, Lindekilde, and Malthaner 2023, 215). That is, varying elements together form different such pathways, such as social relations in their complexity, including how embedded an actor is within a given milieu (Lindekilde, Malthaner, and O’Connor 2019; Malthaner and Lindekilde 2017); mental health, attitudinal, and behavioural factors (Gill et al. 2021; Gill, Horgan, and Deckert 2014); powerful social identities creating in-groups and out-groups (Sageman 2017); biographical factors, including past criminal and gendered violence (McCulloch et al. 2019); and a variety of contingent “trigger” events (Hamm and Spaaij 2017). (Some work ranges over all of the above listed variables – e.g. Lafree et al. (2018) on the criminogenic factors of violent political extremism in general.) So-called triggers or turning points (notions which have been criticised for setting up a false binary – violent/non-violent – where e.g. violence against women is treated as a mere precursor (McCulloch et al. 2019)) – include “moral shocks” and the aforementioned “(action) encouragement cues” (Malthaner and Lindekilde 2017; O’Connor, Lindekilde, and Malthaner 2023; O’Connor, Malthaner, and Lindekilde 2018).

Given this diversity of pathways, it is right to find suspect the way in which some commentators employ the idea of stochastic terrorism, where the inflammatory political speech has the strange compelling power of an imperative. Yet, while informative and useful, we are not quite pushed further to a more structural account of what makes stochastic violence intelligible. For, even at the level of a relational analysis that traces the connections between perpetrators and their milieu, much remains unanalysed that is surely central to stochastic violence.

Thus, what my recommendation here amounts to is keeping centrally in view the ideological substance of stochastic violence, including how this ideology filters through or embeds itself within the mainstream. Others have already well observed the relationship between the far right and the mainstream. Arun Kundnani called for counterterrorist narratives to centre “democratic participation and social solidarity” precisely to “[draw] attention to the wider conditions in European societies which encourage support for far-Right violence” (2012, 30); or, again, when he later portrayed the far-right English Defence League as appropriating mainstream, governmental discourses before giving them “organizational form on the streets” (2014, 241). Likewise, consider Nadya Ali’s argument that “mainstream commentators” are complicit in “producing the anti-immigrant and anti-refugee sentiment” which infused the far-right Thomas Mair’s murder of UK Labour MP Joe Cox (Ali 2020, 591); see also O’Connor, Lindekilde, and Malthaner (2023, 221ff). In addition to such critiques, I suggest that the linguistic device of “stochastic terrorism” (or “stochastic violence”) presents an opportunity to articulate – moreover, I argue it is an opportunity we ought to take up lest the concept be deployed in ways that erase the background by individualising the criminality of the speaker.

My suggestion tracks, to some degree, the practical usage of the stochastic terrorism concept, given its drift in social and academic commentary. While some of the most obvious examples of stochastic terrorism are ones that fit neatly into a schema which abstracts an influential speaker and a persuadable listener (think of the Williams case, e.g. or the “storm”), the term is increasingly used to discuss a broader array of incidents. Emily Bell applies the term to the 2019 massacre in Christchurch, New Zealand, in which Brenton Tarrant attacked a mosque, killing 51 people and injuring 40 more; he had livestreamed part of the incident and publicised it using social media accounts (Bell 2019). Bell notes Tarrant’s “route to terrorism was close to the mainstream, and reflects a constant type of Islamophobia present on news websites with anti-Muslim agendas such as Breitbart and the Daily Mail” (2019). Likewise, consider the also livestreamed massacre in Buffalo, New York (McKinley, Traub, and Closson 2022). The attacker, Payton Gendron, with explicitly racist motives, shot and killed at least 10 victims, mostly black; the text he released to accompany the act was replete with references to the “Great Replacement” conspiracy theory (on which, more below). In the immediate aftermath, even though Gendron cited no particular speaker like Williams had cited Glenn Beck, commentators nonetheless applied the term “stochastic terrorism” – seemingly because it appears inspired by and alike with other notable cases, and because its ideological underpinnings feature prominently a particular conspiracy theory (explored next section) which has been regularly aired on US corporate media (Kayyem 2022; Lavin 2022; Sharlet 2022). Indeed, William Allchorn’s recent book on counter narratives for far-right extremism adopts the term “stochastic far-right terrorism” to cover the Tarrant case and others it inspired in a “cumulative (or … ‘stochastic’) moment” (2022, 18–19).

It seems, therefore, that “stochastic terrorism” has come to be understood as a form of political violence no longer simply represented by an influential speaker who inspires an audience, but also by the larger ideological forces present in mass and social media, and public discourse. Which is to say: the ideology is in the mainstream, which, as Brown, Mondon, and Winter (2021) have argued, adapts itself to far- and extreme-right positioning. Indeed, Tarrant had referenced as an inspiration in his text another mass murder by Anders Breivik in Norway in 2011. Breivik’s own text, in turn, is replete with references to and in some cases extensive quotations from mainstream news journalists who penned anti-immigrant views, including the British Daily Mail journalist Melanie Phillips (Hundal 2011). Despite the “lone wolf” moniker that frequently adorned his name in media write-ups, then, his “attacks were nevertheless inspired by well-known political rhetoric” (Berntzen and Sandberg 2014, 760). And so, as Kundnani argues, it is partly “the significant overlaps of this strand [of far-right ideology] with mainstream views” (2012, 29) which makes such ideology complex. We must contend, then, with a mainstream environment in order to understand some far-right political violence.

Conspiracy and mainstream folk devils

Per the foregoing, there is significant overlap between the ideological inspiration for right-wing extremist violence and public discourse in what is sometimes regarded as the political centre, even as it shifts rightwards. In this connection, it is imperative to consider the mainstreaming of conspiracy theory.

Brenton Tarrant had named the text released alongside his massacre “The Great Replacement”, after the white nationalist conspiracy theory of the same title (Darby 2019).⁷ The replacement theory so-named traces back to the French white nationalist, Renaud Camus; it trades on the claim that elites (hinted or stated to be Jewish) facilitate, through migration, increasing changes in the ethnic and demographic composition of majority white countries. Camus’s claim (itself a variant of older “white genocide” claims) recurs in mainstream politics across Europe, including: in France, by Marine Le Pen, with whom Tarrant sympathised in what he saw as her battle against France’s own great replacement (Rastier 2019); in Germany, by the former SPD politician Thilo Sarrazin, whose very popular 2010 book Deutschland schafft sich ab (Germany Is Doing Away With Itself) discusses similar themes (Heine and Magazzini 2020, 16), and by segments of Alternative für Deutschland, which underwent state surveillance for flirting with these same ideas (Heine and Magazzini 2020, 9); and in the Netherlands, by far-right politician Geert Wilders, who has been direct about believing the conspiracy (Miller 2019). Meanwhile, in the UK, commentator Douglas Murray (Associate Editor of The Spectator) authored a book – The Strange Death of Europe: Immigration, Identity, Islam – espousing essentially the same idea, as economist Jonathan Portes (2021) notes, while continuing to pen articles for centre-right newspapers; and writer Eric Kaufmann’s Whiteshift speaks similarly about demographic change in a way that “treads closely” to the conspiracy theory (Mondon and Winter 2020b, 101), while Kaufman summarises the same topic in the centre-left New Statesman (Kaufmann 2018). Lastly, in the US, (former) Fox News anchor Tucker Carlson had repeatedly promoted the conspiracy theory on his extremely successful show (Bowles 2019; Relman 2021). Worryingly, in fact, some evidence points to approximately one-third of all American citizens believing in the core tenet of the replacement theory (AP-NORC 2022).

Unsurprisingly, therefore, the presence of conspiracy theory in modern authoritarian political violence is not confined to the Breivik or Tarrant cases. In Halle, Germany in 2019, Stefan Balliet attempted a massacre, and while failing to break into the door of a synagogue, he shot and killed a passer-by and man at a kebab shop. Balliet – inspired by Tarrant, livestreaming the attack also – released a text central to which was “a variation on the great replacement narrative” (Heine and Magazzini 2020, 12). Accordingly, a concerning running theme in cases of stochastic violence is the involvement of racist conspiracy theories, while those same theories find mainstream traction. Indeed, finding threads in common weaving these acts together is rarely that coincidental, for they are intended to be “memetic”: attackers reference or adapt memes, some of which run through prior acts of violence (and the texts they accompany), treated as inspirational or memetic in themselves, and some through the forums and sites of discourse of the actors’ online milieus. Such discursive forums include, among others, 4chan or 8chan (later, 8kun)’s /pol/ imageboards, particular subreddits and Discord servers, private Telegram channels, and videogaming media and culture more generally – the livestreaming and use of gamers’ terminology during these events leads some to refer to the violence, rather disturbingly, as a gamified phenomenon (Amarasingam, Argentino, and Macklin 2022; Azani et al. 2020; Macklin 2019; Seymour 2022; Thorleifsson and Düker 2021).

Considering the prominence of conspiracy theory and social media in these kinds of attack, some might think of the online discursive world as a “sort of incubator” for far-right ideology, as the secretary general of Interpol put it (Mekhennet 2020) – perhaps especially in light of the far right’s retreat from more established and open social media platforms to “unmoderated, unregulated fringe platforms” (Azani et al. 2020, 58), therefore occupying something of a private-public space. However, it is easy from here to slip into casting such cultural forces and tendencies, real though they may be, as novel threats acting from the cultural outside-in. By framing the matter as an intrusion of the far or extreme right into a hapless mainstream lacking in agency (see Brown, Mondon, and Winter (2021)), we fail to recall and acknowledge the earlier and ongoing precedents for the wrongful demonising of out-group targets in mainstream news outlets and by governmental representatives (cf. Kundnani (2014, 241)). We fail, that is, to reveal extant authoritarian tendencies at the heart of the modern public square.

To theorise mainstream demonisation, it is useful to reacquaint ourselves with a pair of notions: the “folk devil” and the “moral panics” whipped up against them by institutions of power. Stanley Cohen (1972) first introduced the concepts, today somewhat part of a lay lexicon, in relation to social furores over youth deviance in 1960s Britain. Periodically, argues Cohen, “[a] condition, episode, person or group of persons emerges to become defined as a threat to societal values and interests” (1972, 1). In response, the mass media stylises and stereotypes this folk devil target, before politicians and the commentariat man “the moral barricades” (1972, 1). Cohen’s concepts quickly showed promise in making intelligible the way state and establishment interests pursue their ends through media use and manipulating public attitude. Indeed, Cohen later reflects as much, on “thirty years of moral panics” (Cohen 2002, vii). He outlines their typical criteria: a “soft target” for a folk devil – somebody relatively disempowered; the involvement of “suitable victims” – those that invoke a “could have been me” feeling among the populace; and a developing consensus that the problem runs deep through society, not merely in some fringe actor or event (2002, xii). One category of moral panic he notes has the refugee or asylum seeker as its folk devil: a government- and media-controlled consensus develops to “keep out” refugees, successively portraying them as dishonestly cheating the system, before it finally pushes for stricter criteria of migrant eligibility against a larger “culture of disbelief” around them (2002, xxii).

Even more relevant to understanding this process as authoritarian in nature is the work of cultural theorist Hall et al. (1978). Building on Cohen’s notions, Hall et al. describe the folk devil as “the danger that lurks inside security,” who becomes the bearer of all wrong when that security threatens to unravel (1978, 161). The folk devil they interrogate is the early-mid 1970s Britain “mugger”: as they show, despite the well-known criminal offence of robbery covering instances of theft with force, successive reports by police, law courts, and news media brought into being the mugger as a “new” and continually recognisable, racialised criminal character. The end purpose served by raising the moral panic around the mugger, they argue, is the soliciting of public consent for increasingly realising a law-and-order society: “the themes of race, crime and youth”, which merge into the representation of mugging, ultimately function as a “mechanism for the construction of an authoritarian consensus, a conservative backlash” (1978, viii). This is, then, a process whereby a political and media culture uses myth and mistruth to generate panics (around soft targets) and consent in turn for authoritarian ends. Paying attention to Cohen and Hall et al.’s blueprint, alongside its arguably now plentiful manifestations, brings into view the background social fabric that takes us beyond the individual or even social milieus in analysing instances or types of crime.

The authoritarian deployment of myth of course has an older history still. When Ernst Cassirer reflected in 1945 on the rise of the totalitarian Nazi state, he described myth as precipitating the NSDAP’s rise to power – not by virtue of simply inherited myths or freely imagined ones, but by the newly purposeful creation and direction of myth: “The new political myths do not grow up freely; they are not wild fruits of an exuberant imagination … They are artificial things fabricated by very skilful and cunning artisans” (1946, 182). The novelty is up for question here, given the prominence of myth and propaganda in colonial discourse – both in settler-colonial states, such as the US, and in the metropole of colonies, such as that of the British Empire – and within heavily classed societies. (Cassirer refers, further on, to “savage life” and “primitive societies” (Cassirer 1946, 285–86) even as he compares these caricatures with “modern”, “civilised man” in an attempt to equate their ultimate nature and potential.) Nonetheless, Cassirer was right to emphasise the conscious shaping and use of myth in bringing about authoritarian shifts. What may be different today, after a century’s passing since fascist movements first surfaced, is the fuller repertoire of mythical targets – the “new technique of myth” (1946, 182) (though not entirely new) is now tried-and-tested. So it is not a stretch to connect the presence and spread of the modern replacement conspiracy theory in public discourse with the historical rhetoric characteristic of successive moral panics about immigration in democracies. For Hall et al., the folk devils are, in this sense, pre-constructed and available for more radical right-wing ideologues to exploit for mass appeal.

In this connection, we only have to consider the mass media demonisation of Muslims in the UK and US during and after the War on Terror, especially as the terroristic threat began to be portrayed as no longer external, but as the “threat within”, the process of which Kundnani (2014) illustrates. Likewise, Mary K. Bloodsworth-Lugo and Carmen R. Lugo-Lugo argue that rather than seeing the cultural changes that precipitated the January 6^th “storm” on the Capitol building as stretching only to 2015, “the reorganisation of White supremacist ideologies and groups, along with forms of paranoia and conspiracy theories, have become a driving force within the U.S. mainstream since 11 September 2001” (2023, 185). They offer an effective analysis that shows the intertwining of two decades of discourse about contagion and immigration, yet they also note in passing the colonial history of that linkage (2023, 182–83). (Indeed, per my argument here, we should look this far back, because folk devils, myths, and narratives are so historically embedded.)

Given that popular discourse, culture, and news media may be so well primed to exploit folk devils – to demonise and dehumanise out-group targets – and given the centrality and appeal therefore of racist or reactionary conspiracy theories, we ought not to conceive the threat of stochastic violence as arising from an almost hypnotic power of demagogues at the podium (even paired with unstable “lone wolves” in the audience). Nor is it quite right to expand our focus simply and uncritically to include “the media” as what connects speaker and listener, when our critique only depicts media in turn as a mainstream invaded by extremist ideology. Demagogues have material to work with, a cultural-historical substance dispersed through institutions and public discourse by mainstream institutions of power, and their audiences hear them and their connotations all too well. This, then, is what grounds my embrace of the term “stochastic violence”: with it, we can begin to paint a picture of a larger structure of dehumanisation, wrongful demonisation, propaganda, and ideology that conditions not only the everyday structural and symbolic violences suffered by targeted groups, but specifically also the physical devastation when it manifests in bouts of political violence.

Critical concerns

Notwithstanding the use and significance of “stochastic violence”, we ought to safeguard ourselves too, not least given the warnings that critical studies on terrorism have already unearthed. First, in several formulations of stochastic terrorism, uncertainty is pivotal to the notion: Hamm and Spaaij describe acts of violence which are “statistically predictable but individually unpredictable” (2017, 84), Amman and Meloy say the term “describes a pattern that cannot be predicted precisely but can be analyzed statistically” (2021, 3), and, perhaps most illustratively, Braddock strikes a Rumsfeldian tone: “Because the stochastic terrorist disseminated the inciting message(s) over a wide audience, it is nearly impossible to predict when, where, and by whom a terrorist attack will be performed. However, that an attack will occur is a near certainty” (2020, 225 – emphasis original). In this context, we should be keenly reminded of what Jackson characterises as an “epistemological crisis of counterterrorism”: “an identifiable epistemic posture towards knowledge about, as well as a way of acting towards, the terrorist threat. It manifests itself discursively in the manner in which officials, scholars, pundits and others speak about the threat of terrorism” (2015, 34).

Clearly some of the commentary on stochastic terrorism, from both scholars and pundits, takes up such a posture. It is somewhat given with the very term “stochastic”, given its origin in statistical mathematics denoting a probability that is precisely unpredictable. The main difference would be here the source of the proposed uncertainty: no longer is it due to “shadowy, cunning, adaptive, [and] innovating” terrorists (Jackson 2015, 35), but rather to the unstable and amorphous audience from which a single individual or group will finally, fatally emerge. What Jackson highlighted as consequences of such a crisis, in terms of state and institution security policy, practices, and attitudes, would doubtless not all carry over, but that would not preclude some similar and some novel negative consequences. One which Jackson highlights concerns how the symptoms of terrorism become the focal point of counterterrorism rather than the “deeper roots or causes” (2015, 48). And while I argue the concept of stochastic terrorism in its present uses encourages us to think about causes and conditions, that may not be so for all audiences; another approach – especially one more amenable to state security institutions and policy programmes – would be to focus on very direct causes: namely, speakers and “problematic speech”. One need not be a free speech absolutist for this to prompt anxiety about unwarranted intrusion from states (and non-state actors, considering the outsized influence of some on government) into the domain of speech.

A key characteristic of Jackson’s epistemological crisis is the paranoia among state security establishments and mainstream punditry. However, stochastic terrorism is unlikely to inspire quite this level of paranoia. Part of what drove the fear of “new terrorism” (Spencer 2016) was its being framed as willing to target victims indiscriminately, as having access to mass weapons of nuclear or biochemical destruction, and as waging as a “war” from the outside against a whole way of life. This “newer terrorism” still (Martini and da Silva 2022, 3),⁸ by contrast, appears to target groups which states traditionally treat with less concern or actively oppose: black and minority ethnic groups, the anti-capitalist left, feminist movements, LGBTQIA+ groups, climate protesters, human rights lawyers, and so on.

This broaches a more general concern about the relationship between the status quo and far-right terrorism and political violence. In an introspective on CTS’s ability to engage with far-right subjects, Alice Martini and Raquel da Silva note “the difficulty of focusing on a violence that is aimed at reproducing and maintaining power relations and the status quo” (2022, 6). This point applies as much to mainstream sites of discourse and policy. Consider Ali’s argument (Ali 2020, 587), in the context of the UK’s Prevent counter-radicalisation strategy, that “right-wing extremism is part of a broader normalised continuum of racial violence in Britain” (cf. also Scheper-Hughes and Bourgois (2004)). Indeed, another fruitful framing of the status quo here is that of whiteness, not least as it incorporates “white ignorance” – which in Britain “is tied to the broader project of imperial amnesia” (Ali 2020, 586) – and white supremacy (cf. especially Micieli-Voutsinas and Nguyen (2023, 147) and Bloodsworth-Lugo and Lugo-Lugo (2023)). While much is written and theorised already with these concepts, I think the idea of stochastic violence may help articulate how these larger systems, how the status quo, can filter into authoritarian and far-right political violence.

A last issue to touch on concerns what it is that scholars (including myself) do by re-deploying the language of terrorism, counterterrorism, extremism, and so forth, in the attempt to describe far-right and stochastic political violence. As Jarvis (2022, 14) notes, it is tempting to apply the label “terrorism” to the political violence of the far right as a corrective to previous exclusions of white supremacist violence from discussion and the hyperfocus on a purported connection between Islam and terrorism. Moreover, as he argues, we tend also to essentialise the far right in using this language, without truly taking stock of the “ontological instability” of the phenomena the term supposedly designates. For these reasons, Jarvis (and others for similar reasons before him) recommends we refrain from using the label. We could put it like this: the language of terrorism is “top-down” all the way down. So might the phrase “stochastic terrorism” not just function as another manifest desire for a corrective? There is merit in these arguments. Indeed, I nod towards the term “stochastic violence” partly with such concerns in mind, alongside a worry that the term “terrorism” may obscure vital background structures from view given a tendency of those employing it to individualise violence.

Yet there remains some reason to keep hold of the terrorism label, and not merely as a corrective. A goal of authoritarian violence is to enforce conformity with an ideal or a ruler, circumscribing the “allowable” to a tight area usually coded in terms of race, gender, and sexual orientation (at least). Fear – terror – about what might happen if we contest the allowable by speaking and acting favourably of, or simply just being, the disallowed: that, if anything, is a sure driver of this authoritarian conformity through violence. This line of thought has already been broached, for example by Kundnani: “There are strong arguments for considering all racially motivated violence as a kind of terrorism; … as violence aimed at instilling fear in a population to advance a political cause” (2012, 29). But when we centre a structural understanding of how the individual who enacts political violence comes into being, we incline more to doubt that individual perpetrators are themselves especially motivated by wanting to spread fear and encourage conformity (cf. Sageman 2017, 11). That is the level at which it seems a misnomer (and through which we errantly essentialise, as Jarvis rightly worries). By contrast, at a glance, the motive to martial conformity through fear seems to me far more plausibly attributed to those who would inspire the violence, as well as to the larger ideological structure that charges the authoritarian worldview in the first place.

Notwithstanding that the term “stochastic” opens up potential for the Rumsfeldian exaggerations of our epistemological predicament, it remains significant and useful: it points to the more distant array of factors that coalesce to create the conditions for this form of political violence. Moreover, given the agitations of authoritarian systems – the long-established use of fear as a tool of violent ideology – it might well be premature to entirely do away with the component of terror within the concept. I think some such explanation stands behind why the term has been sought for, expanded, and expressed at ground-level discourse, in fact, and it would be worth seeing this as part of an emancipatory spirit. Accordingly, I hope my response here is congruent with Jarvis’s conciliatory offering to those not fully convinced by deconstructivist aims and methods of critique, which he acknowledges “may not satisfy those committed to a more emancipatory ethic” (2022, 30).

Conclusion

In spite of some concerns over the unclarity of the present uses of the concept “stochastic terrorism”, I both expect it to stay in public and security discourse and to be of some analytical use (though in the latter case, I lean more to the formulation, “stochastic violence”). While the role of intent needs greater clarification, and care should be taken to avoid stumbling into the old epistemological crisis, I hope to have clarified another, overlooked component: wrongful demonisation, and dehumanisation (whose shape the former most often takes). A focus here brings to the surface some of the historical paths of false, misleading, mythical rhetoric – how it plays a structuring role in discursively creating conditions for political violence, a process we can summarise with the notion of stochastic violence. Indeed, I have argued that when we observe the actual use of this nascent terminology, we find that it helps to articulate patterns of authoritarianism in the mainstream of liberal democracies. The term thus encourages us to think also about the causes and conditions of this form of political violence – and what we call terrorism more generally.

In light of the above critical concerns, however, we must still be on the lookout: how might formulations of stochastic terrorism, if they work their way into security parlance, be weaponised against progressive forces or fundamental rights? More generally, when speech becomes the focal point of political and journalistic discourse, we ought to insist on expanding the view to ask what background factors give that speech its significance – how does the mainstream which on the one hand often protests (rightly) against the hateful rhetoric of a demagogue treat the topics of (at least) migration, race, gender, class, and crime? These tasks become more important as the risk Jackson noted before begins to surface again: that counterterrorist discourse and policy has tunnel vision for symptoms and not causes.

Acknowledgements

My thanks to those in my academic department, in particular suggestions from Professor Ralf Poscher, and to Martha Basazienw Kassa for pushing me down this path; and to Dr Roxana Willis, with whom I have shared many vital discussions on these and related topics.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

James Angove

James Angove, Ph.D. is a senior researcher in the Max Planck Institute for the Study of Crime, Security and Law in Freiburg, Germany. With a background and training as an analytic philosopher, his interests span metaphilosophy, later Wittgenstein and Gilbert Ryle, philosophy of language, political and social philosophy, the philosophy of extremism, cultural studies, and terrorism studies. His most recent research project centres on contemporary far-right extremism.

Notes

1. Throughout this article, I employ the convention to use quotation marks or italics for indicating a concept/term, rather than the phenomenon it purports to depict.

2. https://www.dictionary.com/e/what-is-stochastic-terrorism/

3. This example was helpfully suggested to me by an anonymous reviewer, as well as similar examples (in a different context) by Prof. Ralf Poscher.

4. All excerpts from this speech sourced from the Associated Press 2021 transcript: https://apnews.com/article/election-2020-joe-biden-donald-trump-capitol-siege-media-e79eb5164613d6718e9f4502eb471f27.

5. Consider, e.g. the UK Home Office’s “hostile environment” (Mondon and Winter 2020a, 156), and see Hall et al. (1978) on the state role in generating moral panics. See also the uneven, sensationalist reporting on Islamist terror attacks versus other kinds (Kearns, Betus, and Lemieux 2019; Morin 2016; Silva et al. 2020).

6. I thank an anonymous reviewer for pointing out this connection to me.

7. While it is common in media and academia to repeat the self-used term “manifesto” here, I think it rather exalts the text; I borrow “text(s)”, albeit not in explicit reference to “manifestos”, from Harmonie Toros (2022, 229).

8. I am not here endorsing the new/old distinction.