Dissertation

Social Processes and Collective Mobilization in the Online Extreme Right

Overview

My dissertation investigates the mobilizing capacity of the online extreme right — a movement characterized by physical dispersion, organizational fragmentation, and demographic breadth, yet exhibiting a striking level of strategic convergence and aesthetic unity. Drawing on literatures from political science, social psychology, extremism studies, and social computing, I develop and test a theory of how collective mobilization occurs among decentralized online extremist communities despite the absence of formal organizational hierarchy, fluid and transient memberships, and geographic dispersion.

The Puzzle

Conventional understandings of social movements and terrorism suggest that cohesion, organizational unity, and well-defined systems of resource mobilization are important — if not necessary — to achieve movement goals. Yet the online extreme right has produced a wave of mass-casualty political violence with perpetrators who are seemingly diffused throughout the world, often know few fellow believers offline, and plan attacks independently with support and guidance coming only from online communities. The proscription of specific groups and increasing law enforcement pressure have not managed to disrupt ongoing apocalyptic-oriented violence.

This pattern of “templatized” violence has developed within and across leaderless online communities like 4chan, Telegram, and Gab. Perpetrators reproduce elements from prior attacks — tactical choices, manifestos, symbols — in ways that suggest powerful but informal social processes at work.

Central Questions:

Given the physical dispersion, organizational fragmentation, and demographic breadth of the online extreme right, from where does its mobilizing capacity emerge?
What drives the contagion of the form and intended function of apocalyptic violence?
Is it possible to empirically describe the latent relational infrastructure of a movement as decentralized as this?

Theoretical Contribution

I propose that extreme-right collectivities produce violent attacks when they establish social processes such as the widespread adoption of apocalypticism, the crystallization of mobilizing artifacts (like texts), and the establishment of a “canon” of attackers. Integrating findings from the literatures of political violence, contentious politics, and social computing, I identify where conventional expectations of movements and mobilization may miss more complex dynamics in online contexts.

I use digital archival materials to qualitatively build a theory of collective mobilization in the online extreme right and identify testable implications of this theory across three empirical chapters.

Empirical Chapters

Chapter 2: “The Happening. This Time It’s Real”: Apocalyptic Rhetoric on 4chan Following Mass Casualty Attacks

4chan’s Politically Incorrect (/pol/) image board has been linked to numerous hate crimes, physical violence, and acts of mass-casualty terrorism, yet the mechanics of radicalization and mobilization on /pol/ are still largely unexplored. This chapter investigates the relationship between offline attacks and online communities of extreme actors by exploring how apocalypticism on /pol/ responds to real-world acts of mass-casualty violence.

Data: Archived posts from 4chan’s /pol/ board
Methods: Transformer-based language classifier applied to identify apocalyptic rhetoric; intervention analysis on resulting time series
Key Finding: Online apocalypticism significantly increases in the immediate aftermath of mass-casualty terrorist attacks, providing evidence for feedback loops between online communities and violent attacks

Chapter 3: Resurrecting James Mason: The Development of “Siegism” on Iron March

This chapter investigates the crystallization of ideology and culture on Iron March, a neo-fascist web forum active between 2011–2017. I theorize that Iron March developed a cohesive and sustainable collective identity through its establishment of a “memory artifact” in the form of a forum-published edition of James Mason’s Siege. The adoption of Siege was foundational to Iron March’s capacity to produce anti-democratic militancy, including revolutionary fascist paramilitary groups and terrorist collectives.

Data: Complete archive of Iron March forum posts
Methods: Interrupted time-series analysis, network contagion analysis
Argument: Communities can contribute to offline mobilization despite no formal organizational hierarchy, fluid and transient membership, and geographic dispersion through the crystallization of mobilizing artifacts

Chapter 4: “Canonization” of Violent Attackers in the Extreme-Right Apocalyptic Movement

This chapter argues that symbolic and tactical contagion in decentralized social movements is facilitated through a cyclical process of collective commemoration and canonization. Beginning with the observation that extreme-right communities refer to their heroized attackers as “Saints,” I develop a measure to describe the extent to which these communities canonize particular attackers using NLP techniques applied to eleven years of 4chan /pol/ posts.

Data: Large-scale 4chan /pol/ archive; novel dataset of every attempted or successful mass-casualty attack in OECD countries over the past 20 years (drawing from ACLED, START’s Global Terrorism Database, the RTV dataset, and other sources)
Methods: Natural language processing, regression analysis with novel “Sainthood” measure as dependent variable
Expected Findings: Attacks that are more severe and more explicitly connected to the extreme-right apocalyptic movement become canonized at higher rates; attackers that adopt the tactics and aesthetics of past “Saints” are more likely to become Saints themselves

Broader Contributions

Beyond its empirical findings, this dissertation makes contributions relevant to computational social science and the study of political violence more broadly:

Theoretical: Understanding collective processes can significantly improve our understanding of political violence in online and leaderless contexts
Methodological: Combining NLP, network analysis, and causal inference in a unified framework to study decentralized movements
Policy: Efforts to mitigate political violence should take into account “background social conditions” — the communities from which violent attackers emerge, and which connect many attacks into a constantly growing web of violence

Student Research Involvement

Each dissertation chapter offers opportunities for undergraduate research assistants to contribute meaningfully:

Chapter 2 (Apocalyptic Rhetoric): Students can assist with annotation of 4chan posts for training the transformer-based classifier, validation of time-series coding, and data preparation
Chapter 3 (Siegism on Iron March): Students can contribute to qualitative coding of forum posts, network data construction, and analysis of ideological diffusion patterns
Chapter 4 (Canonization): Students can help build the novel mass-casualty attack dataset, code attack characteristics, and assist with variable construction for the regression analysis

I have supervised 45+ undergraduate and graduate research assistants across my research career, and I actively seek to integrate students into all stages of the research process.

Committee

Co-Chairs:

Andrew Q. Philips, Department of Political Science, CU Boulder
Jennifer Fitzgerald, Department of Political Science, CU Boulder

Members:

Alexandra Siegel, Department of Political Science, CU Boulder
Brian Keegan, Department of Information Science, CU Boulder (external advisor)

Timeline

Milestone	Target Date
Prospectus Defense	Completed
Chapter 2 Draft	Summer 2026
Chapter 3 Draft	Fall 2026
Chapter 4 Draft	Spring 2027
Defense	May 2027

For questions about my dissertation research, contact alex.newhouse@colorado.edu.