Category: Blog

The recent workshop held by doctoral candidates Camilla Gissel, Heidi Campana-Piva and Violette Mens at the STS-CH conference in Zurich entitled “Holding things together? Change, continuity, critique” (from the 10^th to 12^th of September) began with a provocation: a deepfake video of Kamala Harris circulating online. The clip, shared on Elon Musk’s platform X, had garnered nearly a million likes and tens of thousands of comments. What seemed at first like obvious parody—her voice and image altered to mock her campaign messaging—was, for many viewers, indistinguishable from reality.

This set the stage for a lively discussion not just about deepfakes themselves, but about how technology interacts with political radicalization in an era where truth is increasingly fragile.

When a Joke Stops Being a Joke

One participant raised the question: is a parody video like this really a deepfake, or just satire in digital form? After all, political cartoons have long exaggerated politicians’ flaws for comic effect. Yet others pushed back. Unlike cartoons, which signal their artifice, deepfakes thrive in ambiguity. Some lines in the Harris video were things she had actually said, others were fabrications. That blurring—between critique, parody, and falsehood—creates a puzzle for viewers.

Even when it is “obvious” to some, not everyone has the same interpretive tools. As one participant noted, people already committed to conspiracy thinking (e.g. flat-) can believe almost anything if it confirms their worldview. Deepfakes exploit that cognitive vulnerability.

The Politics of Doubt

The group then shifted to a related danger: once deepfakes exist, politicians can weaponize them to dismiss inconvenient truths. One example shared was of Trump brushing off journalists’ questions about suspicious activity at the White House by declaring “It’s AI. It’s fake.” Whether or not it was fake became irrelevant—what mattered was the ability to cast doubt.

This erosion of shared reality is not accidental. Participants pointed to Trump’s thousands of documented lies in office and the way constant confusion about truth can destabilize citizens. Destabilization breeds fear, and fear drives people to seek stability—often in the arms of authoritarian leaders who promise certainty. In this way, deepfakes are not just tools of deception; they are accelerants in the cycle of radicalization.

Radicalization: More Than a Label

The conversation broadened to the term “radicalization” itself. Too often, the word functions as a blunt political instrument. Governments use it to stigmatize dissent, lumping together jihadists, eco-activists, and radical feminists under the same umbrella. By labeling groups “radicalized,” states can justify surveillance, repression, or even violence.

Several participants argued that radicalization is relational: it doesn’t happen in isolation but through interaction between groups and the state. In France, for example, jihadist violence has fueled harsher policing, which in turn produces resentment and further radicalization. This feedback loop shows radicalization as a dynamic process, not simply a personal pathology.

Others raised the Overton Window: the shifting boundary of what society considers politically acceptable. As mainstream politics drift rightward, advocating for basic human rights can suddenly be branded “radical left.” The term becomes a moving target, often manipulated to discredit opponents rather than to explain genuine extremism.

Beyond Extremes: Who Gets to Define Radical?

An especially striking thread was the comparison between the far right and far left. Media often presents them as mirror images—two extremes equally dangerous. Yet, as some participants noted, the comparison is misleading. The far right frequently undermines democratic norms, while much of the “radical left” remains engaged in democratic processes, calling for rights and reforms rather than authoritarian control.

Academics, too, have tended to study far-right and Islamist extremism while neglecting other forms of radical politics. This selective focus reveals how research agendas themselves are shaped by political pressures, such as the aftermath of terror attacks.

Why This Matters

Deepfakes may seem like a technological novelty, but as the workshop discussion made clear, they are deeply entwined with broader political struggles. They blur the line between fact and fiction, fuel cycles of distrust, and give political actors new tools to label, demonize, and radicalize.

But they also force us to ask hard questions about the words we use. If “radicalization” is applied too broadly, it loses analytical value and becomes little more than a weapon of discourse. And if truth itself becomes negotiable, then the fight is not just about politics, but about the very possibility of shared reality.

The workshop’s starting point—a fake Kamala Harris video—was more than just a gimmick. It was a reminder that in our current moment, what matters is not simply whether something is “true” or “false,” but how technologies of persuasion and doubt are reshaping the terrain of democracy itself.

INTRODUCTION

Most authors agree that the Eurabia conspiracy theory started with the publication of the book entitled Eurabia – The Euro-Arab Axis, in 2005, by a French author named Giséle Littman, but published under the pseudonym Bat Ye’or. The text states that “ever since the early 1970s, the European Union was secretly conspiring with the Arab League to bring about a ‘Eurabia’ on the continent” (Bergmann 2021: 39). In 2011, another French author called Renaud Camus published a book entitled The Great Replacement, that “argued that European civilisation and identity was at risk of being subsumed by mass migration, especially from Muslim countries” (Ibid, 37). These books introduced the “fear of cultural subversion” that is characteristic of this conspiracy theory.

Eurabia also presuppose three states: First, a paradisical past when Europe was only populated by Caucasians (at least in the interpretations of these conspiracy theorists). Then, a present danger which configures a fall from paradise; white people are disappearing due to immigration and low birth rates of ‘native’ Europeans. And lastly, redemption, the envisioning of a better future; plans for making Europe return to its supposed cultural, ethnic, and religious roots.

My research aims to semiotically analyse the messages from a white supremacist Telegram group, with the help of digital humanities methodologies, seeking language patterns that can potentially assist in the codification of cultural meanings and in the formation of these anti-Muslim ideological clusters on Telegram. The main contributions of my work to the filed of Semiotics is the incorporation of computational tools in the analysis of text in large-scale (allowing for both data size and data depth), and the contribution to Digital Humanities is to go beyond only the detection of conspiracy theories in online content but towards structural analysis without sacrificing context, which is a big problem in the field of computational tools applied to humanities and social sciences research nowadays.

MATERIALS & METHODS

Unfortunately, even though the Telegram channel itself is public, I am unable to share its name as it is sensitive information protected by the GDPR. The data that I obtained from the channel was the textual non-pictorial content of messages sent from its administrators to the channel’s subscribers (which amount to more than 22 thousand accounts). The messages were collected from January 1^st, 2023, to December 31^st, 2024, totalling more than 4 thousand messages, varying in length.

The method of analysis is still being developed. We are applying what is called Semantic Annotation with Linguistic Inquiry Word Count using the layout of FrameNet (a lexical database being developed at the International Computer Science Institute in Berkeley since 1997).

Simply put, text annotation is adding a tag to a text excerpt. Basically, we are teaching the computer to understand which terms and expressions have similar semantics and share common contexts so that they can be represented in space close to each other or have the same representation. This way, meaning is approximated so that words can be represented in a lower dimensional space.

The first step in the annotation process was the entity selection – that is, choosing specific instances (people and organizations) that are interesting subjects of discourse. Basically, I was looking for specific texts that potentially discuss certain topics that are of interest for analysis.

The next step was the definition of the taxonomy, focusing on the core-elements of a conspiracy theory. Basically this means deciding on the specific categories to annotate the texts with. Initially, we accessed the FrameNet database and found that they do not have an annotated dataset for “conspiracy” – which is excellent, since this is what we are trying to make. Instead, they give the “closest” results which are: Collaboration and Offense. Using these 2 as examples, I developed the Frame Index for Conspiracy Theory. After making a list of interesting entities and having the well-defined taxonomy, we generated random samples for annotation.

A scheme of the developed Fame Index can be found on the image below:

Obviously, each text will, most of the time, present only a few of these categories, which is fine. If the software can learn to flag the excerpts that have 2 or 3 of these tags, they can go into the “to be analysed by a human” box. This could be a way to use computation to make human analysis more efficient. By separating the “useful” extracts for analysis and displaying them with the pre-identified tags, then, a deeper discourse analysis can be carried out by the semiotician.

We are still in the annotation process which means I do not have the results from the computational analysis yet. But so far, I’m confident that this is a good way to help scholars to quickly gain insights from these huge datasets. This white supremacist channel does not only disseminate Eurabia conspiracy theories, but by teaching the computer to understand and summarize what are the out-groups, in-groups, evil plans, for each text, one can easily paint the picture of the main structure of a conspiracy theory narrative, allowing scholars to not just identify their presence in a dataset, but also understand their main elements and how they are related, without having to go through the whole textual content, which would be quite time-consuming, not to mention emotionally exhausting due to the pernicious character of these messages’ content. Since we’re still developing this, I cannot say with 100% certainty that it will work, but I believe in the relevance of trying.

DISCUSSION

Now, to close up, I would like to discuss the Religion problem, since this is the European Academy of Religion congress. The last time I presented this case study at the Peace Research Institute Frankfurt, I had interesting feedback. People were asking me “Why are you treating this conspiracy theory as a religious one? It’s not religious, it’s political”. So, I thought that this conference would be a good opportunity to present my take and see what do other scholars from religious studies think of this issue.

According to the literature, in Eurabia and Great Replacement discourses, ‘Islam’ is associated with “evil, crime and barbarism”, as well as other “harmful characteristics and ideological markers that enhance polarised, emotional and simplifying visions of social reality” (Gualda 2021: 57). It is “typically represented as backwards, fanatic and violent”, as well as a totalitarian political doctrine (Dyrendal 2020: 374), while Muslims themselves “are generally portrayed as a homogeneous group of violent and authoritative religious fundamentalists” (Bergmann 2021: 42). Muslim individuals are seen as “mere executors of a religiously based, collective will” and, consequently, since Islam is itself seen as fundamentalist in nature, “every believer will be made to follow its radical version” (Dyrendal 2020:  374). In this sense, the idea of ‘Islam’ is seen as being a uniting factor for all Muslims, that unites them “in a common plan for domination” (Ibid).

In this sense, the “Eurabia conspiracy theory has often become entangled in a more general opposition to immigration” (Bergmann 2021: 40), connected to political statements of failed integration (Ekman 2022: 1128) – which are based around the notion that Western societies are homogeneous, and that Muslims and other migrants are unable to integrate into them (Gualda 2021; Ekman 2022) – or to the notion that “incorporation of diversity, multiculturalism or other elements of Islam or the Muslim world into [Western] culture” will mean the total collapse of society, which will become a colony of Islam (Gualda 2021: 61-62). In other words, the arrival of “new norms, habits and customs brought by the foreign population […] could influence the disappearance of one’s own culture” (Ibid), turning immigration into an ‘invasion’ that threatens people’s identity.

So, what we have here, is first of all, a very problematic conflation between Arab world and Muslim world. The attribute of religious identity based on ethnic and geopolitic identity is a problem in itself. But let us take a quick step back.

Asbjørn Dyrendal (2020) describes these three kinds of dynamics that can be used to express the relationships between conspiracy theories and religion. The first one, conspiracy theories in religion, relate mostly to authority and power, since they are usually employed to delegitimize those that are seen as enemies of a certain religious group. The second one, conspiracy theory as religion, regards the idea that conspiracy theories are replacing religion by exerting its functions in a now more secularized society. This notion can be questioned, since it is first of all not possible to state that we have more conspiracy theories today than during a time when religious adherence was supposedly stronger, and also because “religion is usually not negatively correlated with conspiracy beliefs”, suggesting the two go hand-in-hand, rather than one replacing the other (Dyrendal 2020: 373). Instead of thinking of conspiracy theory as a substitute of religion, we may think of the ways in which conspiracy theory can be seen as a form of religion, given the status of both religion and conspiracy theories as alternative or counter-knowledge, as well as how they both organise collective identities on the basis of in-group and out-group.

But I want to focus on the last one, conspiracy theories about religion, or how conspiracy theories are formed regarding certain religious groups. Eurabia is an ethno-religious myth. As a researcher, I am aware of the complexities in these narratives and I obviously don’t buy this conflation between Arab and Muslim, but it is a matter of how the analysed discourse is constructed – the Emic point of view. To the endorsers of Eurabia discourse, there is no distinction, they don’t fear Christian Arabs. I would argue most of them don’t even know there is such a thing as Christian Arabs. They fear what they think Islam is (since they are also ignorant of the complexities of Islam itself). And of course there is another dimension to this issue which is the fact that conspiracy theories are not completely misaligned with the contexts that favour certain representations. These notions about the Arab world, Islam, and Muslims are not constructed in a vacuum. Media representations of Islam contribute to the construction of stereotypes in conspiracy theories as well. 

FINAL REMARKS

So, in conclusion, the Eurabia conspiracy theory was brought firmly into the political mainstream by the financial crisis of 2008 and later the refugee crisis of 2015 (Bergmann 2021: 48-49). Nowadays, it is common to find political leaders who propagate Great Replacement and Eurabia conspiracy theories in the mainstream media (Ekman 2022: 1127). As we see such Islamophobic racist discourses become more popular, we also see them become normalized, especially across new media platforms such as Telegram. This means research needs to adapt to these new contexts, and digital humanities tools become invaluable for these efforts.

REFERENCES

Bergmann, E. (2021). The Eurabia conspiracy theory. In: Önnerfors, A., & Krouwel, A. (Eds.), Europe: Continent of Conspiracies: Conspiracy Theories in and about Europe. Routledge, 36-53.

Dyrendal, A. (2020). Conspiracy Theory and Religion. In: Butter, M., Knight, P. (Eds.), Routledge Handbook of Conspiracy Theories. Routledge, 304-316.

Ekman, M. (2022). The great replacement: Strategic mainstreaming of far-right conspiracy claims. Convergence, 28(4), 1127-1143.

Gualda, E. (2021). Metaphors of Invasion, Imagining Europe as Endangered by Islamisation. In: Önnerfors, A., & Krouwel, A. (Eds.), Europe: Continent of Conspiracies: Conspiracy Theories in and about Europe. Routledge, 54-75.

Introduction

The Polish series The Hooligan (Kibic) ^[1], released on Netflix in January 2025, has drawn significant attention for its raw and gripping portrayal of violence, loyalty, and identity. In five intense episodes, it follows the story of Kuba, a teenager pulled into the ruthless world of ultra-football supporters. While the series is of poor cinematic interest, it exceeds the cliché of the brawl-seeking delinquent and offers a more nuanced exploration of the radicalization process, showing that this phenomenon often goes far beyond mere ideological motivation.

A story of youth, belonging, and fracture

The Hooligan follows Kuba, a young man torn between his mother’s love, the return of his ex-convict father, and the allure of power and belonging offered by the local ultra-group, the Kosynierzy. Violence is ever-present, but it’s rarely senseless. It operates within a framework of recognition, territory, shared values, and above all, identity. We slowly see Kuba, the kind son, hard-working student and caring brother sink into violence when he joins the fitness gym held by the Kosynierzy. What started as a will to follow onto his father’s footsteps and prove his strength, becomes a little help with the gym work, quickly transformed into a drug dealer side job and escalating into several settling of scores, many ending with the death of loved ones. 

Kuba doesn’t become violent by nature. He becomes violent because it’s a means of carving out a place for himself, of fulfilling needs unmet by his home or social life. This violence is also supported by an allegiance system to the ultra-group as well as a system of value that puts personal honor center stage, even to the detriment of life. 

This logic aligns closely with the Good Lives Model, a theoretical tool increasingly used to understand paths to radicalization.

The Good Lives Model: a humanistic view of risk

The Good Lives Model (GLM), created by Ward and Maruna (2007) “is premised on the idea that the best way to reduce risk is to help individuals live better lives, not simply to teach them to avoid offending. It emphasizes the importance of human agency, strengths, and values in promoting pro-social behavior.”^[2]. This rehabilitation model was developed to better understand why individuals engage in extreme or deviant behavior primarily in the context of offender rehabilitation, especially with individuals convicted of sexual or violent crimes, but its application has expanded to fields like radicalization and terrorism prevention. It is based on the idea that every person seeks to attain “primary goods” such as life meaning, self-control, relationships, autonomy, or belonging. When these fundamental needs are frustrated or inaccessible through conventional means, some may turn to violence or extremism to meet them.

From this perspective, radicalization is not merely a theoretical allegiance to a cause. It is a response to a rupture in meaning or recognition. It can appear, albeit toxically, as a perceived solution to deep malaise. The GLM emphasizes not dangerousness but the potential for rebuilding a good, dignified, and meaningful life.

Radicalization in “The Hooligan”: an identity-based process

In The Hooligan, Kuba’s motivations for joining the ultras are never rooted in ideological commitment. What drives him is the need to be recognized by his father, to earn respect, to be part of a powerful group, and to have a clear role in a world that offers him none. The series shows that this radicalization is built through layers: dysfunctional family, financial struggles, trauma, and the need for protection.

The show also illustrates how radical groups exploit these personal vulnerabilities to lure young people in search of meaning. Zyga, the ultra-leader, embodies a paternal and manipulative figure who provides structure, purpose, and a mission. Within this dynamic, violence becomes a language, a proof of commitment, a rite of passage.

A call for understanding, not complacency

What The Hooligan succeeds in doing is showing that radicalization is not about monsters to be eliminated, but about human beings to be understood. By viewing Kuba’s story through the GLM lens, we better grasp how violent paths take root. And more importantly, we understand that the exit from radicalism requires the reconstruction of meaning, healthy connections, and a hopeful outlook.

Far from glorifying or minimizing violence, The Hooligan raises an essential question: what are we trying to fill when we turn to extreme behaviors? And more importantly: what can we offer, as a society, so that these quests find more humane answers?

---

^[1] https://www.netflix.com/TITle/81618034

^[2] Ward, T., & Maruna, S. (2007). Rehabilitation: Beyond the Risk Paradigm, Routledge.

Evaluation fulfils a variety of roles in the field of preventing and countering violent extremism (P/CVE), which often involves a range of non-state and civil society actors in primary, secondary and tertiary prevention activities. Against the backdrop of evaluation’s historical role in public policy, where evaluation often informed budget decisions (cf. Derlien, 1998), it is often associated with a legitimising function. This is particularly true in a highly competitive funding landscape, in which projects often compete for limited funds (Malet 2021; KN:IX 2020). When state funding bodies request evaluations, this can create the impression that evaluations are primarily control instruments (Sivenbring and Andersson Malmros 2019; cf. Treischl and Wolbring 2020).

However, evaluation can also be emancipatory. Self-evaluation in particular has the potential to professionalise the practice of preventing and countering violent extremism and to establish a stronger professional self-image (von Berg et al. 2024, 216). This opportunity contrasts with a reality in which the topic of evaluation has only increasingly come into focus in the past years (cf. Uhl and Kattein 2024; Bressan et al. 2024) yet it is still often criticised as inadequate in terms of scope and quality (cf. Feddes and Gallucci 2015; van Hemert et al. 2014). There are many reasons for this: On the one hand, there is frequently a lack of evaluation skills at an individual level, and, on the other hand, there is often a lack of institutionalised procedures and conditions at a structural level. The latter is currently being investigated as part of a country-comparative dissertation within the VORTEX doctoral network under the provisional title ‘Evaluation and Quality Management in the Field of Preventing and Countering Violent Extremism – Twelve European Countries in Comparison’. However, there are also starting points at an organisational level to make evaluations more practicable: Appropriate structures and processes can, for example, demystify external evaluations and make the potential of self-evaluations more valuable and achievable. This blog post outlines three such levers at the organisational level: a systematic and impact-oriented practice, a monitoring system and a guideline for the case of evaluation within the organisation.

Systematic and impact-oriented practice

A P/CVE practice that is systematically structured and impact-oriented is much easier to evaluate than practice in which impact assumptions, intervention goals and success criteria have not been defined in advance. The development of a theory of change for the overall strategy is a fundamental first step, as it creates a common understanding of how and why certain measures should contribute to the achievement of objectives (INDEED 2023, 16).

However, it is also crucial to address the level of specific measures: Impact assumptions and intervention objectives should be precisely formulated and operationalised with measurable success criteria. In the evaluation of federally funded counselling centres in Germany, Karliczek et al. (2023) propose a model that differentiates objectives in disengagement work. The nine-field matrix structures goals both in terms of the reference level – i.e. whether they relate to the radicalised person themselves, to their interfaces with the environment or to the social environment – and in terms of the level of disengagement work, which distinguishes between pragmatic aspects, socio-affective-emotional dynamics and ideological-normative convictions.

If such intervention goals have been systematically developed and concretised using success criteria, counsellors can use them to reflect on the results of their work with clients, for example through measuring how their clients fare on each of them before and after an intervention (Karliczek et al. 2023, 106). A practice that works on this basis makes it much easier to document the impact of disengagement work in a comparable way across many cases. However, this is not about rigid standardisation, but about structured flexibility: in other words, a framework concept that guides professional action but does not restrict individual casework. One example of such a systematic procedure is the social diagnostic model adapted by the Violence Prevention Network e. V. for disengagement work. It combines a holistic analysis of cases with needs-oriented interventions and creates a sound basis for making impact both visible and assessable (von Berg et al. 2024). Importantly, it leaves flexibility to counsellors within the model so that they can decide which measures are necessary and appropriate in a given context. 

Establishing a monitoring system

Such a systematic approach naturally lends itself to monitoring, i.e. the ongoing, systematic recording of data. Applied to the aforementioned possibilities of measuring the results of disengagement work, this means regularly documenting whether a positive change has taken place for a client in relation to a specific intervention goal based on the specified impact assumptions and intervention goals. If we move away from the counselling example and towards trainings with multipliers on how to handle situations related to violent extremism, a key question might be whether participants feel more knowledgeable after the training. Tracking these self-assessments over time is essential for meaningful evaluation. While monitoring focuses on ongoing data collection, evaluation can go a step further: it may ask how effective the intervention is in achieving its intended goals and makes a judgement about its impact. In doing so, evaluation can build on the data collected through monitoring and analyse it to assess overall effectiveness (Junk 2021).

Being prepared for evaluation

Even if an organisation structures its work systematically and monitoring is firmly anchored in practice, conducting a self-evaluation – and especially an external evaluation – can be perceived as overwhelming. If the organisation is not sufficiently prepared, there is a risk that the potential of an impact-oriented approach to monitoring will remain unused. It is therefore advisable to think about responsibilities and coordination processes within the organisation at an early stage. Klöckner et al. (2021), for example, describe the benefits of an evaluation working group within civil society organisations focused on supporting clients with disengaging from violent extremist groups. Such a group can represent a recurring early exchange format within the organisation in preparation for an external evaluation. With regard to an upcoming external evaluation, such a working group offers a protected framework in which expectations, fears and specific needs can be openly addressed. In this way, the organisation can collect and communicate its needs and objectives at an early stage (expectation management) and discuss the possibilities of practical feasibility and already look at how evaluation results can be transferred into practice (cf. Klöckner et al. 2021, 7). The overarching goal of the evaluation working group is to strengthen the organisation’s identification with the evaluation process and thus ensure the greatest benefit of the evaluation for the organisation (INDEED 2023, 28). However, such an evaluation working group is also suitable for the preparation and coordination of self-evaluations. Here, too, it can be a recurring structure that accompanies the process of harnessing monitoring data for evaluation. 

This blogpost sought to explore how P/CVE organisations can better harness the potential of evaluation by embedding three structural levers into their work: First, developing systematic, impact-oriented practices, second, establishing monitoring systems to track progress over time, and third, preparing for evaluation through internal coordination processes such as a working group on evaluation. These strategies help make both self-evaluation and external evaluation more meaningful and manageable, thus transforming evaluation from a perceived burden into a valuable learning opportunity.

Sources

Bressan, Sarah, Sophie Ebbecke, and Lotta Rahlf. 2024. ‘How Do We Know What Works in Preventing Violent Extremism? Evidence and Trends in Evaluation from 14 Countries’. Berlin: GPPi; PrEval (PRIF). https://gppi.net/assets/BressanEbbeckeRahlf_How-Do-We-Know-What-Works-in-Preventing-Violent-Extremism_2024_final.pdf.

Derlien, Hans-Ulrich. 1998. ‘Le Développement Des Évaluations Dans Un Contexte International’. In Politiques Publiques: Évaluation, 7–11. Paris: Economica.

Feddes, Allard R., and Marcello Gallucci. 2015. ‘A Literature Review on Methodology Used in Evaluating Effects of Preventive and De-Radicalisation Interventions’. Journal for Deradicalization, no. 5 (December), 1–27.

Hemert, Dianne van, Helma van den Berg, Tony van Vliet, Maaike Roelofs, and Mirjam Huis in ’t Veld. 2014. ‘Synthesis Report on the State-of-the-Art in Evaluating the Effectiveness of Counter-Violent Extremism Interventions’. Deliverable 2.2. IMPACT Europe.

INDEED. 2023. ‘How to Design PVE/CVE and De-Radicalisation Initiatives and Evaluations According to the Principles of Evidence-Based Practice’. INDEED Consortium. https://home-affairs.ec.europa.eu/networks/eu-knowledge-hub-prevention-radicalisation/welcome-package/learning-resources/indeed-e-guidebook-2-how-design-pvecve-and-de-radicalisation-initiatives-and-evaluations-according_en.

Junk, Julian. 2021. Quality Management of P/CVE Interventions in Secondary and Tertiary Prevention: Overview and First Steps in Implementing Monitoring and Reporting. Radicalisation Awareness Network. https://home-affairs.ec.europa.eu/system/files/2021-12/ran_ad-hoc_quality_management_of_p-cve_interventions_122021_en.pdf.

Karliczek, Kari-Maria, Vivienne Ohlenforst, Dorte Schaffranke, Dennis Walkenhorst, und Juliane Kanitz. 2023. Evaluation bundesfinanzierter Beratungs-stellen: Abschlussbericht der Evaluation der Beratungsstellen zur Distanzierung und Deradikalisierung vom islamistischen Extremismus. Beiträge zu Migration und Integration, Band 12. Nürnberg: Bundesamt für Migration und Flüchtlinge. Junk, Julian. 2021. Quality Management of P/CVE Interventions in Secondary and Tertiary Prevention: Overview and First Steps in Implementing Monitoring and Reporting. Radicalisation Awareness Network. https://home-affairs.ec.europa.eu/system/files/2021-12/ran_ad-hoc_quality_management_of_p-cve_interventions_122021_en.pdf.

Klöckner, Mona, Svetla Koynova, Johanna Liebich, and Lisa Neef. 2021. ‘Erfahrungen aus der Evaluationsplanung eines Aussteigerprogramms. Voraussetzungen für Wirksamkeitserfassung in der tertiären Extremismusprävention’. PRIF Report 6. Frankfurt am Main: PrEval Consortium. https://www.prif.org/fileadmin/Daten/Publikationen/Prif_Reports/2021/PRIF0621_barrierefrei.pdf  

KN:IX. 2020. ‘Kompetenznetzwerk “Islamistischer Extremismus” (KN:IX) – Herausforderungen, Bedarfe und Trends im Themenfeld’. Berlin: Kompetenznetzwerk Islamistischer Extremismus. https://kn-ix.de/wp-content/uploads/2021/02/KNIX-Report-2020.pdf.

Malet, David. 2021. ‘Countering Violent Extremism: Assessment in Theory and Practice’. Journal of Policing, Intelligence and Counter Terrorism 16 (1): 58–74. https://doi.org/10.1080/18335330.2021.1889017.

Sivenbring, Jennie, and Robin Andersson Malmros. 2019. Mixing Logics: Multiagency Approaches for Countering Violent Extremism. Göteborg: Segerstedinstitutet, Göteborgs Universitet. https://www.gu.se/sites/default/files/2020-03/1764750_korrekt-versionmixing-logics_digital_korrekt.pdf.

Treischl, Edgar, Tobias Wolbring. 2020. Wirkungsevaluation. Grundlagen, Standards, Beispiele, Weinheim/Basel: Beltz. 

Uhl, Andreas, and Ian Kattein. 2024. ‘Monitoring von Evaluationskapazitäten in der Extremismusprävention, Demokratieförderung und Politischen Bildung’. In PrEval Monitor: PrEval Zukunftswerkstätten. https://preval.hsfk.de/fileadmin/PrEval/PrEval_Monitor_2024_engl..pdf

Von Berg, Annika, Dennis Walkenhorst, Gloriett Kargl, and Maximilian Ruf. 2023. Soziale Diagnostik in der Extremismusprävention – Diagnose, Fallverstehen, Intervention und Wirkungsmessung. Ideologie und Gewalt – Schriften zur Deradikalisierung. Wiesbaden: Springer Fachmedien Wiesbaden. https://doi.org/10.1007/978-3-658-42427-5.

VORTEX DC Heidi Campana Piva and her colleagues are offering a workshop on language processing and ai tools applied to research in the humanities and social sciences at the University of Turin in October 2025.

This workshop offers PhD students a practical opportunity to learn how to apply a few Natural Language Processing (NLP) tools to their research. Despite its current popularity, Artificial Intelligence and NLP products are still exclusive tools used by few – only a limited number of large companies have the resources at their disposal to invest in effective NLP solutions. With that in mind, this workshop aims to make AI-powered applications a more accessible asset.

During our two encounters, we will attempt to demystify AI tools, in order to empower young researchers to benefit from its potential without being unaware of its implications in society, aiming to make AI processes more human-oriented and transparent, also contributing to better understanding of how AI-based services reach their decisions.

The workshop will take place within the framework of the MSCA project VORTEX, using its research as basis for learning – that is, in order to learn how to apply NLP tools, the workshop will be presenting the issues of online radicalization, asking participants to carry out a short case-study. Students will learn annotation of corpora and be introduced to text analytics. At the closing session, participants will have the opportunity to reflect upon their achieved results, as well as how they each can make use of the newly learned tools in their own research.

Practical information:

INSTRUCTORS

Heidi Campana Piva – MSCA PhD fellow, Department of Philosophy and Educational Sciences, UniTo.

Marco Antonio Stranisci – Research fellow, Computer Science Department, UniTo.

Supervision: Massimo LEONE, Full Professor, Department of Philosophy and Educational Sciences, UniTo.

SCHEDULE

First Introductory Session: 15 October 2025, from 9:00 to 13:00.

Follow-up Closing Session: 12 November 2025, from 9:00 to 13:00.

DURATION: 8 hours

MODALITY: In-person only Workshop

LOCATION: Auditorium Quazza, Palazzo Nuovo, Via Sant’Ottavio 20, Torino.

LANGUAGE: ENGLISH

TARGET

PhD Candidates of any research area who have interest in working with computational linguistics.

ENROLMENT

To participate in this workshop, it is necessary to register by filling in the dedicated Google Form. 

This WS has limited capacity (50 seats, first come first served basis).

Registration is open until spots are available and anyway not later than October 3.

Given the limited number of seats available, we kindly ask you to register to the WS only if you are certain that you can participate and, should impossibility to participate arise after you registered, to let us know via email (contact at the bottom of this document).

Participants are advised to bring their own computers.

CONTACTS INFORMATION:

heidi.campanapiva@unito.it

In moments of global crisis, narratives move faster than facts. As states exchange missiles and accusations, and headlines race to assign blame, a central question often goes unasked: how do we come to understand military violence? For most of it is not through direct experience, but through stories. And those stories are not neutral. 

Narratives are not just matter-of-fact commentaries; they are structuring forces. As Philip Smith argues in “why war?”, narrative forms are essential to the very organization of political (and military) action. They make complex realities easy to understand by reducing them into cause, effect, villains, and heroes. In times of deep conflict, these narrative frames become even more rigid. Simplification is not merely incidental to violence; it is what makes violence possible.

Smith argues that war narratives often adopt an apocalyptic genre, casting large scale (usually military) violence as both unavoidable and cathartic. In the current escalation involving Iran, Israel and the United States this genre is no longer just theoretical abstractions. Israeli Prime Minister Benjamin Netanyahu recently stated that Israel’s airstrikes had eliminated “two immediate existential threats” referring to Iran’s nuclear facilities. This statement exemplifies the genre’s moral coding: Iran becomes more than an adversary – Iran becomes a threat to (Israeli) civilization and violence is framed as necessary. It is either them or us. 

This juxtaposition is not confined to Israeli rhetoric. A month prior to the stacks United States President Trump referred to Iran as “the most destructive force in the Middle East” and warned them against an Iranian continuation of “Chaos and Terror”. The reference to Iran causing terror is most likely not a coincidence but ties in with what we usually refer to as ‘the war on terror’ a powerful reference that has deep emotional memories in western countries in general and the United States in particular. In such a tale, the space for proportionality collapses. Violence is not framed as one policy option among many, but as a moral imperative.

These narratives are not mere stylistic choices; they help shape public tolerance for violence, what leaders justify, and determine how the moral lines are drawn. Peace, in this frame, is not the result of genuine dialogue, but part of a narrative orchestration. Negotiation of peace by and large becomes a symbolic gesture that is not performed to resolve conflict, but to legitimize strategic power already in full and deadly motion. 

We see narrative power operating in real time. Claims of ceasefire violations are made before independent evidence is available. Missile launches are reported and then denied. Harm to civilians is referred to selectively, if at all. (Try counting the (lack of) international articles concerning how the escalation affects Iranian civilians). Meanwhile, terms such as ‘retaliation’, ‘defense’ and ‘response’. These are terms which pre-shapes public understanding in positioning some forms of violence as legitimate and others as deviant. Who is perceived to have started the conflict can depend less on chronology and more on narrative framing. 

The narrative structure of violence often mirrors apocalyptic myth more than reportage. A cycle of threat and safety. A cycle of peace shattered, and order restored. It is exactly this apocalyptic tale that makes continued violence appear not only reasonable but a necessity. Once a narrative becomes dominant it tends to absorb contradictions rather than fall apart: “we are defending ourselves” becomes the interpretive filter. New facts don’t necessarily disprove it instead they get reinterpreted to fit the filtered lens no matter how logically flawed they are. 

This is not to say that all actors are the same, or that truth is unattainable (at least in theory). But truth in escalations like these is always mediated. As someone who studies political violence my focus should not only be on what happened, but how what happened is made meaningful. Which voices are heard loud and clear, which silences are sanctioned, who are demanded to speak when they would rather not, and which scripts are recycled?

The current escalation between Israel, the United States and Iran with its contested ceasefires and choreographed press conferences, offer a painful but important reminder: war is always narrated. In this gap between action and understanding lies the space where power does something of its most enduring work. 

Germany’s professional practice in the fields of extremism prevention, democracy promotion, and civic education is characterised by a notably high number and diversity of knowledge networks. These networks aim to connect civil society and state actors, facilitate professional exchange, and thereby strengthen the overall professionalisation of the field. In a European study on national and regional P/CVE networks, Dzhekova et al. (2024) highlight the unique German context:

‘In some countries, such as Germany, the multitude of actors involved in this work at state (province) level and the diversity of P/CVE programmes require stronger networking and exchange of experience between practitioners (but also with research and policy) to ensure the professionalisation of the field.’

During the 2020–2024 funding period of the federal programme ‘Live Democracy’, a total of fourteen so-called competence centres were funded, including the Competence Network on Right-Wing Extremism (KompRex) and the Competence Centre for the Prevention of Islamism (KN:IX). In the new funding period which began in 2025, these centres turned into the cooperation networks Competent in Right-Wing Extremism Prevention(KompRex, n.d.) and KN:IX connect (KN:IX connect, n.d.). Additionally, these phenomenon-specific networks are supplemented by others that deal specifically with counselling practice for exit and distancing/deradicalisation work, mobile counselling and counselling for victims and others affected by violent extremism. 

The the very same federal programme has seen the building of a comprehensive evaluation infrastructure. While ‘scientific company’ of individual funded projects and the programme as a whole has been an integral part of the federal programme from the beginning, as well as in previous programmes, the third funding period is accompanied by an evaluation network, consisting of six research institutes (BMI, n.d.).

A significant contribution to professionalising German prevention practice in recent years regarding evaluation, importantly also beyond such programme structures as  ‘Live democracy!’, has been the research and transfer project PrEval Zukunfswerkstätten. For the first time, it systematically surveyed evaluation capacities both nationally (Uhl et al. 2022) and internationally, with a view to deriving recommendations for Germany (Baykal et al. 2021; Bressan et al. 2024). It also identified the needs of practitioners regarding evaluation (Koynova et al. 2022), presented relevant methods (e.g. Klemm and Strobl 2024), and conducted pilot studies (e.g. Kindlinger et al. 2025). In addition, PrEval developed various support formats to foster methodological capacity-building and facilitate knowledge transfer within the field. Internationally, the project is considered exemplary for advancing the professionalisation of evaluation and quality assurance in the fields of extremism prevention, democracy promotion, and civic education.

One particular study analysed the need for support formats among practitioners and how they could be integrated into existing knowledge networks (Koynova et al. 2024). A visibility analysis showed that although various networks exist in the fields of extremism prevention, democracy promotion and civic education, there are hardly any publicly visible references to evaluation or quality assurance (Ibid., 26; cf. Tschöp et al. 2023). Nevertheless, in a follow-up survey, which primarily reached evaluators, six networks were named as relevant networks dealing with the topics of evaluation and quality assurance (ibid.). However, there were specific preferences regarding the type of network in which such topics were discussed. The authors found that ‘while practitioners (…) tend to describe access to thematically relevant networks in which they also exchange information on other issues in their fields of work, participants from academia or evaluation (…) predominantly describe access to networks focussing on evaluation and quality assurance, with which they are already familiar due to their focus of work […]’ (Tschöp et al. 2023, 8; translated from German; Koynova et al. 2024, 27). 

While no explicit reference to evaluation was found in the networks most frequently used by practitioners, it was assumed that internal exchanges on these topics do occur. Yet, the needs assessment clearly demonstrated practitioners’ need for exchange on evaluation and quality assurance, as well as for concrete support for evaluation projects (Koynova et al. 2022). To fill this gap, the authors of the PrEval study recommended to systematically strengthen the visibility and discoverability of relevant evaluation expertise within existing networks (Koynova et al., 2024, 27-28). Drawing on both the needs assessment and the visibility analysis of existing networks, several support formats were developed that can either be integrated into existing knowledge networks or serve to strengthen connections between them (ibid., 33). One such format is the ongoing counselling series ‘Evaluation and Quality Assurance’, currently piloted by the Violence Prevention Network and i-unito. The initiative is designed to support civil society organisations in developing and implementing evaluation and quality assurance processes. The programme offers modular support in areas such as designing impact models, defining indicators, developing data collection instruments, and analysing evaluation data (Giel 2025).

Another complementary project dealt with the conception of a helpdesk for evaluation and quality assurance in the fields of extremism prevention, democracy promotion and civic education in Germany. Surveying for needs among practitioners revealed that some would like to link such a service to established civil society organisations or networks in order to avoid any competitive dynamics (Ruf et al. 2024, 13). It was found that these networks already function as trustworthy contact points (ibid., 20) and that their use could help to avoid duplicate structures in the already complex prevention landscape (ibid., 27).

With the conclusion of the PrEval project at the end of 2025, a comprehensive body of knowledge is now available, both on the needs of practitioners and other stakeholders, and on how to implement practice-oriented support formats. Moreover, there are early experiences with piloting these formats. Key ‘products’ of the project such as the ideas and pilots of an evaluation database, low-threshold support formats, and the counselling series represent essential tools that can advance the professionalisation of evaluation and quality assurance in the field. 

The next crucial step is to maintain these formats and embed them institutionally. Since PrEval opened spaces to discuss evaluation beyond specific funding structures, existing knowledge networks lend themselves well for this purpose. While networks such as KN:IX are financed by ‘Live Democracy!’, they are not limited to projects of this funding structure. Moreover, their bottom-up character are well suited to discuss sensitive topics such as evaluation in an environment that is perceived as ‘safe’. Research on the institutionalisation of evaluation shows that such a process cannot succeed without a broader system of professionalisation (Meyer, Stockmann and Taube, 2020; cf. Stockmann and Meyer, 2016). Core components of such a system include structured training opportunities, peer learning formats, and shared professional standards. As several PrEval studies have highlighted, linking emerging support formats to established thematic knowledge networks could offer a practical pathway to strengthening a ‘system of professionalisation’. 

However, the success of this transfer largely depends on the capacities and resources available within those networks. The expertise developed within PrEval since 2020 is therefore more than just an output of the project – it is a critical enabler for future institutionalisation efforts. In other European countries, where civil society plays a less prominent role in prevention, there are typically fewer knowledge networks in the P/CVE field and related areas, as well as fewer research projects that explicitly research evaluation and invest in capacity building. Some isolated initiatives exist, such as institutionalised helpdesks in the Netherlands or best practice exchange formats in Denmark, but these are usually state-led and primarily target municipalities. By contrast, Germany’s knowledge networks provide alternative avenues to embed exchange and support formats for evaluation within civil society practice. The challenge now is to fully leverage this potential.

Sources

Baykal, Asena, Sarah Bressan, Julia Friedrich, Giulia Pasquali, Philipp Rotmann, and Marie Wagner. 2021. ‘Evaluating P/CVE: Institutional Structures in International Comparison’. GPPi; PrEval (PRIF). https://preval.hsfk.de/fileadmin/HSFK_Preval/GPPi_2021_Extremismuspraevention_evaluieren_Institutionelle_Strukturen_i….pdf.

BMI. n.d. ‘Evaluation’. https://www.demokratie-leben.de/dl/programm/evaluation.

Bressan, Sarah, Sophie Ebbecke, and Lotta Rahlf. 2024. ‘How Do We Know What Works in Preventing Violent Extremism? Evidence and Trends in Evaluation from 14 Countries’. Berlin: GPPi; PrEval (PRIF). https://gppi.net/assets/BressanEbbeckeRahlf_How-Do-We-Know-What-Works-in-Preventing-Violent-Extremism_2024_final.pdf.

Dzhekova, Rositsa, Maximilian Ruf, Sophie Kuijper, Margareta Wetchy, and David Tschöp. 2024. ‘National and Regional Networks Focusing on P/CVE across the EU’. Radicalisation Awareness Network. https://home-affairs.ec.europa.eu/document/download/b70fc126-5856-498b-a5c6-abb5d88299b3_en?filename=ran_national_regional_networks_focusing_on_P-CVE_across_EU_en.pdf.

Giel, Susanne. 2025. ‘Beratungsreihe Evaluation und Qualitätssicherung’. Frankfurt am Main: PrEval Consortium. https://preval.hsfk.de/fileadmin/PrEval/PrEval-Beratungsreihe/PrEval-Beratungsreihe_Evaluation_und_Qualit%C3%A4tssicherung_gesamt.pdf.

Kindlinger, Marcus, Lucy Huschle, and Hermann Josef Abs. 2025. ‘Entwicklung eines integrierten Modells und Selbsteinschätzungsinstruments für Digital Citizenship Literacy’. PrEval Studie 1. Frankfurt am Main: PrEval Consortium. https://preval.hsfk.de/fileadmin/PrEval/PrEval_Studie_1_2025.pdf.

Klemm, Jana, and Rainer Strobl. 2024. ‘Wirkungsmodelle und ihr Potenzial für Evaluation und Qualitätssicherung in der Demokratieförderung’. PrEval Expertise. Frankfurt am Main: PrEval Consortium. https://preval.hsfk.de/fileadmin/PrEval/PrEval_Expertise_01_2024.pdf.

KN:IX connect. n.d. ‘Islamistische Radikalisierung geschieht nicht im luftleeren Raum’. KN:IX Connect. n.d. https://kn-ix.de/.

Kober, Marcus, Lotta Rahlf, Mikhail Logvinov, Julian Junk, Helle Becker, and Marcus Kindlinger. 2024. ‘Zukunftswerkstatt “Evaluationsdatenbank”: Bericht’. Frankfurt am Main: PrEval Consortium. https://preval.hsfk.de/fileadmin/PrEval/Bericht_ZW_Evaluationsdatenbank.pdf.

KompRex. n.d. ‘Das Netzwerk’. Kompetenz Netzwerk Rechtsextremismus Prävention. n.d. https://kompetenznetzwerk-rechtsextremismuspraevention.de/das-netzwerk/.

Koynova, Svetla, Lina Hartmann, Moritz Lorenz, David Tschöp. 2024. ‚Bericht der Zukunftswerkstatt Wissensnetzwerke‘. Frankfurt am Main: PrEval Consortium. https://preval.hsfk.de/fileadmin/PrEval/Bericht_ZW_Wissensnetzwerke.pdf.

Koynova, Svetla, Alina Mönig, Matthias Quent, and Vivienne Ohlenforst. 2022. ‘Monitoring, Evaluation und Lernen: Erfahrungen und Bedarfe der Fachpraxis in der Prävention von Rechtsextremismus und Islamismus’. 7. Frankfurt am Main: Leibniz-Institut Hessische Stiftung Friedens- und Konfliktforschung (HSFK). https://www.hsfk.de/fileadmin/HSFK/hsfk_publikationen/PRIF_Report_7_2022_barrierefrei.pdf.

Meyer, Wolfgang, Reinhard Stockmann, and Lena Taube. 2020. ‘The Institutionalisation of Evaluation: Theoretical Background, Analytical Concept and Methods’. In The Institutionalisation of Evaluation in Europe, edited by Wolfgang Meyer, Reinhard Stockmann, and Lena Taube, 3–36. Cham: Palgrave Macmillan. https://doi.org/10.1007/978-3-030-32284-7_1.

Ruf, Maximilian, David Tschöp, Janusz Biene-Clément, Moritz Lorenz, Alina Mönig, Toqa Hilal. 2024. ‚Bericht der Zukunftswerkstatt Helpdesk.‘ Frankfurt am Main: PrEval Consortium. https://preval.hsfk.de/fileadmin/PrEval/Bericht_ZW_Helpdesk.pdf.

Stockmann, Reinhard, and Wolfgang Meyer. 2016. ‘The Future of Evaluation: Global Trends, New Challenges and Shared Perspectives’. In The Future of Evalaution: Global Trends, New Challenges, Shared Perspectives, edited by Reinhard Stockmann and Wolfgang Meyer, 9–19. London: Palgrave Macmillan. https://doi.org/10.1057/9781137376374.

Tschöp, David, Emma van Heeswijk, Lina Hartmann, Benedikt Büchsenschütz. 2023. ‚Auf der Suche nach Expertise. Die Schwierigkeiten bei der Auffindbarkeit von Wissensnetzwerken und Evaluationsexpertise in der Demokratieförderung, Extremismusprävention und politischen Bildung.‘ PrEval Working Paper 1. Frankfurt am Main: PrEval Consortium. https://preval.hsfk.de/fileadmin/PrEval/Working_Paper_AP5_21.09..pdf

Uhl, Andreas, Manuela Freiheit, Benjamin Zeibig, and Andreas Zick. 2022. ‘Evaluationskapazitäten im Bereich der Extremismusprävention und der politischen Bildung in Deutschland’. 9. Frankfurt am Main: Leibniz-Institut Hessische Stiftung Friedens- und Konfliktforschung (HSFK). https://doi.org/10.48809/prifrep2209.

INTRODUCTION

White Genocide conspiracy theories surround the notion that native white people of predominantly white countries are being dis/replaced with alien people of colour as a result of a hostile alliance between domestic and foreign political-economic elites. This idea has been traced all the way back to pre-World War II, with antisemitic conspiracy theories narrating the existence of a Jewish plot to destroy Europe through miscegenation, having deep historical roots in French nationalism (Davis 2025), especially with the book The Uprooted (1897) by Maurice Barrès. In the XIX century, it was common for nationalist politicians to compare France’s low birth-rate with the high birth-rates of East-Asian countries of that time (Anderson 2014). From such negative comparisons, the fear of Asian mass-migration arose. More recently, scholars often cite the book The Great Replacement (2011) by another French author called Renaud Camus as a relevant contribution to the conspiracy theory, this time targeting Muslims and other migrants from North Africa and the Middle East (Bergmann 2021).

In general, White Genocide conspiracy theories highlight the fear of cultural subversion, accusing a domestic internal elite of betraying the native white people into the hands of an external evil. Intrinsically tied with ani-immigration discourses, these conspiracy theories presuppose three states: First, a paradisical past; the Good Old Days when Europe/North America were only populated by Caucasians (at least in the interpretations of these conspiracy theorists). Then, a present danger; white people are disappearing due to immigration and low birth rates. And lastly, the envisioning of a better future; plans for making Europe/North America return to their supposed cultural, ethnic, and religious roots.

In concise terms, conspiracy theory can be defined as a representation in the form of a narrative that seeks to explain a determined circumstance as being the result of a secret plan implemented by a morally evil group of people that, if left unstopped, will lead to catastrophe (adapted from Birchall 2006 and Önnerfors 2021). From this, the definition of the White Genocide conspiracy theory may be that there is scheme by political and economic elites in predominantly white countries to cause the extinction of what are perceived as native white populations through the promotion of miscegenation, multicultural and racial integration policies, mass immigration, low fertility rates and abortion of native people, and organised violence (adapted from Jackson 2015 and Davis 2025).

The aim of this research is to identify language patterns that can potentially assist in the semiotic modelling / codification of cultural meanings and in the formation of white-supremacist ideological clusters on social media. The research question that guides this study is: What types of signs, texts, and codes structure the White Genocide conspiracy theory?

MATERIALS AND METHODS

Unfortunately, even though the Telegram channel itself is public, I am unable to share its name as it is sensitive information protected by the European General Data Protection Regulation (GDPR). The data that I obtained from the channel was the textual (non-pictorial) content of messages sent from its administrator to the channel’s subscribers (which amount to more than 22 thousand accounts). The messages were collected from January 1^st, 2023, to December 31^st, 2024, totalling more than 4 thousand messages, varying in length.

The method of analysis is still being developed. I am working in partnership with the Computational Linguistics department of the University of Turin to refine available tools for the analysis of right-wing conspiracy theories. We are applying what is called Semantic Annotation with Linguistic Inquiry Word Count (LIWC) using the vocabulary from the Moral Foundations Dictionary (MFD). I will explain shortly how this works.

ANNOTATION PIPELINE

The first step was the entity selection – that is, choosing specific instances (people and organizations) that are interesting subjects of discourse. Basically, I was looking for specific texts potentially discuss certain topics of interest for analysis. The software is able to identify people and organizations on itself, however it does not understand that, for instance, UN and U.N. are the same thing. Therefore, when I was selecting the subjects of interest, I had to also compile the different ways that the same person/org could show up in the dataset, as you can see in this screenshot I took. 

On the left, there are the outputs that the software automatically generated and on the right there are my list of interesting entities and how they may appear. After making a list of interesting entities, we generated random samples for annotation. Three sets (for 3 annotators) of 400 messages (each) were randomly assembled from the total of all texts containing the selected entities. This way, each annotator would be able to read and annotate a set of 400 messages. But what does annotating actually mean?

Simply put, text annotation is adding a tag to a text excerpt. Basically, we are teaching the computer to understand which terms and expressions have similar semantics and share common contexts so that they can be represented in space close to each other or have the same representation. This way, meaning is approximated so that words can be represented in a lower dimensional space.

The idea is that, once the dataset is annotated, LIWC will able to – for a given text input – return output lists of words falling into each category. These are the meaningful that we are mapping:

• pronouns – first-person / third-person / singular / plural
• ingroup-outgroup language – us vs. them; native vs. alien/foreign; white identity vs. racial resentment
• concerns- work, leisure, home, money, death…
• time orientation – past, present, future + sentiment polarity in relation to time – negative view of the present / positive view of the past? What about the future?
• affective processes – positive emotions (joy, safety, celebratory feelings, trust…); negative (anxiety, anger, disgust, sadness, fear); and neutral (anticipation, surprise…)

With those interests in mind, the third step was the definition of the taxonomy, that is, deciding on specific categories to annotate the texts with. These are the categories that I came up with, believing them to be useful for analysis of conspiracy theory structure:

• enemy / out-group / them
• victims
• in-group / us
• evil plan / evil deed
• mis/disinformation
• glorified past / reactionary
• fear of the future
• danger of the present

For instance, in a given text we have “concerned citizens” as the in-group. Then, we have the sentence that states that “activism today quickly turns violent” in contrast with the idea that this was not the case with activism in the past (danger of the present). The sentence “issues ignored by the mainstream” indicate that these problems are secret, conferring the conspiratorial aspect to the text. On another text, the ‘evil plan / evil deed’ category is labelled onto a sentence stating “loss of 1.3 million more jobs while foreign-born workers gained more than 1.2 million.” Also the expression “erasing your existence” has been found to be very telling of the White Genocide conspiracy theory in these texts.

FINAL REMARKS

But so far I’m confident that this is a good way to help scholars to quickly gain insights from these huge datasets. This white supremacist channel does not only disseminate White Genocide conspiracy theories, but by teaching the computer to understand and summarize what are the out-groups, in-groups, victims, evil plans, present danger, and fear of the future for each text, one can easily paint the picture of the main structure of a conspiracy theory narrative, allowing scholars to not just identify their presence in a dataset, but also understand their main elements and how they are related, without having to go through the whole textual content. Since we’re still developing this, I cannot say with 100% certainty that it will work, but I believe in the relevance of trying. And if it doesn’t work, at least I’ll have a map of why it didn’t work and how future research may develop better ways to do this same thing.

REFERENCES

Anderson, M. C. (2014). Regeneration Through Empire: French Pronatalists and Colonial Settlement in the Third Republic. University of Nebraska Press. p. 25.

Bergmann, E. (2021). The Eurabia conspiracy theory. In: Europe: Continent of Conspiracies: Conspiracy Theories in and about Europe. Routledge, 36-53

Davis, M. (2025). Violence as method: the “white replacement”, “white genocide”, and “Eurabia” conspiracy theories and the biopolitics of networked violence. Ethnic and Racial Studies, 48(3), 426-446.

Jackson, P. (2015). ‘White genocide’: Postwar fascism and the ideological value of evoking existential conflicts. In The Routledge history of genocide (pp. 207-226). Routledge.

Birchall, C. (2006). Knowledge goes pop: From conspiracy theory to gossip. Berg Publishers.

Önnerfors, A. (2021). Conspiracy theories and COVID-19: The mechanisms behind a rapidly growing societal challenge. Myndigheten för Samhällsskydd och Beredskap.

Research covering the role of religion in the right-wing populist discourse often adopts the frame of “instrumentalization”, hijacking, exploitation, or even weaponization to explain the strategic use of religion for political gain (Giorgi, 2022). For instance, Samaras (2025) analyses and compares how three far-right Greek parties – Greek Solution, Victory, and Voice of Reason – draw on religious narratives to advance exclusionary politics between 2019 and 2024. He goes on to claim weaponization of Greek Orthodox Christianity by the far-right parties to adopt nationalism, xenophobia, and anti-progressive ideologies, often framed as resistance to so-called “woke” movements (Cammaerts, 2022). 

In 2016, a book called: Saving the people: how populists hijack religion (Marzouki et al.) analyzed mobilization of religion by the right-wing populists in 8 different countries: United States, Britain, France, Italy, Austria, the Netherlands, Poland and Israel. The authors concluded that for populists and their supporters, religion is not as much about belief as it is about belonging. But what if some of them see themselves as true believers—and reject this interpretation? 

This question has been taken up by scholars like Alberta Giorgi (2022), who examined how Salvini’s supporters themselves understand their religious identity. She challenged the common view that radical-right populist parties “instrumentalize” religion. She looked closely at what supporters of Matteo Salvini had to say—and found that they strongly reclaimed the legitimacy of their Christian identity. They openly opposed the “religion of the elites,” criticized the Catholic Church’s hierarchy, and promoted a form of “people’s religion” instead.

As Giorgi points out, this reclaiming of religion—positioned against traditional religious authorities and institutions—combined with an emphasis on the cultural side of Christianity, reflects a broader transformation in how religion functions today. In post-secular societies, both religion and politics have undergone processes of de-institutionalization and disintermediation. This means people are moving away from relying on established religious leaders and institutions and gaining more direct control over how they interpret and practice their faith.

From this perspective, political leaders aren’t simply “hijacking” religion—because religion is already active outside of traditional institutions. Supporters of radical-right populist parties aren’t just passive victims of political manipulation—they often share the same personal and cultural understanding of religion. In other words, religion isn’t so much “hijacked” as it is freed from institutional constraints and reinterpreted in new ways. 

This insight connects to a wider debate today about who has the right to speak for and about religion. According to Giorgi’s analysis, Salvini’s supporters reject the idea that religion is being politicized or polarized against their interests. Instead, they claim the legitimacy of being religious while voting for a party whose positions don’t always align with those of institutional religious authorities. The key argument is that the Catholic Church and official religious leaders are not the only legitimate representatives of Catholicism or religious faith.

While Giorgi’s analysis restores agency to social actors and challenges simplistic views, the concept of “instrumentalization” remains useful for understanding populist leaders’ strategies and messages. It helps explain how political actors navigate political and discursive opportunities and how discourse functions in the political arena. Moreover, it highlights the contradictions between the policies radical right parties promote and the religious values they often invoke in various ways.

Interestingly, the relationship between right-wing populist parties and Church leaders varies significantly across countries. For example, while Matteo Salvini’s supporters openly criticize “church elites” in Italy and promote a form of popular religion distinct from institutional Catholicism, in Poland influential religious figures like Tadeusz Rydzyk, a powerful priest who runs a business empire centered around the media outlets Radio Maryja and TV Trwam, maintain close ties with the populist Law and Justice (PiS) party. In fact, the party has been accused of “buying” Rydzyk’s support by using public funds to finance political agitation through his media outlets. This contrast shows how the entanglement of religion and populism can take very different forms depending on national and cultural contexts. 

In conclusion, the relationship between right-wing populist politics and religion is complex and multi-layered. Moving beyond the idea of simple “instrumentalization” allows us to better understand how religion functions as a living, evolving part of identity and culture, shaped both by leaders and supporters alike. Recognizing these dynamics—and the variations across different countries—can help us understand the nuanced role religion plays in today’s political landscape.

References

Cammaerts, B. (2022). The abnormalisation of social justice: The ‘anti-woke culture war’discourse in the UK. Discourse & Society,33(6), 730-743. 

Giorgi, A. (2022). Hijack or release? On the heuristic limits of the frame of instrumentalization of religion for discussing the entanglements of populism, religion, and gender. Identities, 29(4), 483-499. 

Marzouki, N., McDonnell, D., & Roy, O. (2016). Saving the people: How populists hijack religion. Hurst. 

Samaras, G. (2025). Weaponising Religion in Greece Post Economic Crisis: An Analysis of Far-Right Political Discourse from 2019 to 2024. Journal of Empirical Theology, 1-28. https://doi.org/https://doi.org/10.1163/15709256-20240018

Certain topics naturally bring together not only academics but also policymakers and professionals who work on them. Preventing and countering violent extremism (P/CVE) and, by extension also terrorism, is one example, where research intersects with real-world challenges faced by law enforcement, civil society actors, and policymakers. These professionals, who often enough work in their own filter bubbles, have much to gain from networking. Yet, opportunities for cross-sectoral mutual learning and networking often remain limited due to structural as well as professional hurdles. This blog post describes the opportunities, but also the challenges, of these rare occasions of coming together. 

The most common format for meetings are thematic conferences, which are open to researchers as well as representatives from politics and practice. One example of this is the recently held 6th annual conference of Europol’s European Counter Terrorism Centre (ECTC) in April 2025, titled “Rank and Files or Memes and Likes? Terrorist Organisations, Groups, Networks.” As the title suggests, the conference centred on violent extremists’ and terrorists’ activities in the online sphere – a particularly complex and evolving challenge for policymakers and law enforcement agencies but also a topic that has caught increased scholarly attention. This is a commonality of such networking conferences: They tend to focus on those topics that are felt as the most pressing and emerging challenges in countering violent extremism and terrorism. Networking events, therefore, serve as a marketplace for informational demand and supply, ensuring that discussions remain relevant and forward-looking. In this marketplace, policymakers and practitioners seek insights on current and future threats and face challenges on dealing with them. Many so-called national action plans, strategic policy documents on preventing and countering violent extremism, emphasise the need for evidence-based decision-making. Even when policymakers do not immediately recognise the relevance of academic research, exposure to new insights can have indirect but significant effects on policy formulation. Likewise, practitioners, whether from law enforcement or civil society, can share their experiences and provide feedback on how existing P/CVE policies or research resonates with their work. Lastly, for researchers, such a conference offers the opportunity to influence policy and practice with their research findings, either by offering new insights or by providing critical perspectives.

Two examples from the ECTC conference can be highlighted since the programme of the conference was published and both projects are publicly described. One example for research influencing policy and practice is the VEOMAP research project funded by the Polish National Agency for Academic Exchange and realised at the ITSTIME research centre in Milan, which drew on open-source intelligence and used social network analysis to explore the pro-Islamic State (IS) information ecosystem and its evolution on the surface web (July 2023-March 2024) (see Lakomy 2024). By showing the structure, among others the quite centralised nature of this ecosystem, such research may resonate with similar findings from law enforcement agencies, go beyond them or inspire practice methodologically. A different example, showing the possibility of sharing experiences with P/CVE policy, albeit without going too much into the detail of this process, is the proscription of the Terrorgram collective as terrorist organisation on 26 April 2024 in the United Kingdom (see Home Office 2024). Sharing the steps and implications of this progress as well as perspectives on its effectiveness allowed policymakers and law enforcement agents from other countries to contemplate whether similar efforts could be realised in their respective national contexts.

While the picture painted here sounds promising, it is not without its problems in reality. Two interlinked problems will be discussed here. Firstly, professional differences between the actors often stand in the way of learning from one another and fruitful dialogue. Policymakers and law enforcement agents are often said to ‘think’ differently from prevention actors, as the former focus strongly on eliminating security risks and the latter have a broader understanding of prevention that also looks at the strengths of individuals and society against violent extremism. It seems counterintuitive at first, but to bridge these differences, which is necessary to enable a fruitful exchange and learning from each other, a lot of dialogue and trust is needed. One conference alone will not be able to create this. Instead, regular formats are needed that enable interpersonal relationships and a ‘meeting again’, an important basis for building trust. Only in such a setting is it conceivable that P/CVE practitioners in particular, some of whom have a special relationship with P/CVE policymakers due dependency relationships caused by public funding programmes, will be encouraged to share their experiences and criticism. Only when trust has been established do researchers get the impression that they are met with honest interest and open ears. However, structural bottlenecks in the P/CVE field, especially a lack of more stable funding, mean that exchange formats tend to appear sporadically rather than being permanent. Without naming examples here, there are some of these kinds of exchange formats that started promisingly and then fizzled out due to domestic political changes in the country that funded the event or due to the expiry of the one-off funding period. Others have been good spaces for exchange for years, where both new and the same people have been able to meet again and again, and which are now at risk of dissolution due to an administrative change, and it remains unclear to what extent previous formats will be continued. Those who want to secure the benefits of these exchange formats would do well to ensure that they are not a one-off event or that their future is uncertain, but that they take place regularly and thus allow some trust-building in the P/CVE and counterterrorism field.