Repeated and incontrovertible collective action failure leads to protester disengagement and radicalisation

Abstract

Protest is ubiquitous in the 21st Century and the people who participate in such movements do so because they seek to bring about social change. However, social change takes time and involves repeated interactions between individual protesters, social movements and the authorities to whom they appeal for change. These complexities of time and scale have frustrated efforts to isolate the conditions that foster an enduring movement, on the one hand, and the adoption of more radical (unconventional, unacceptable) tactics on the other. Here, we present a novel, theoretically informed and empirically evidenced, agent-based model of collective action that provides a unified framework to address these dual challenges. We model ~10,000 iterations within a simulated society and show that where an authority is responsive, and protesters can (cognitively and/or socially) contest the failure of their movement, a moderate conventional movement prevails. Conversely, where an authority repeatedly and incontrovertibly fails the movement, the population disengages but becomes radicalised (latent radicalism). This latter finding, whereby the whole population is disengaged but prepared to use radical methods to bring about social change, likely reflects the febrile pre-cursor state to sudden, revolutionary change. Results highlight the potential for simulations to reveal emergent, as-yet under-theorized, phenomena.

Figures (13)

• Figure 1: Schematic overview of four phases of the DIMESim model. First, an Authority broadcasts to all agents a common success or failure signal relating to their demands (1). The signal is received by each agent who can (2) individually re-frame the signal from failure to success (e.g. shown by Bill and Cissy), and, in a third step, (3) collectively re-frame based on the success/failure interpretations of neighbours in the social network. Finally (4), based on the perceived signal and the agent's DIME variable states, the agent will make a decision about whether to act (active or latent), in either a conventional, innovating, or radical direction. In the schematic example, all agents have received a failure signal (red arrow). Abby interprets the failure signal as a failure and does not individually or collectively re-frame that failure as success (red exclamation). Bill collectively re-frames failure as success (green tick). Cissy individually re-frames failure to perceive the outcome as a success (green tick).
• Figure 2: The emergence of very different outcomes under conditions of responsive vs. intransigent authorities, individual and collective re-framing (a),(b): Network examples ($n=50$) at $t\in\{1,10^2,10^4\}$ and, (c)-(f): aggregate timeseries plots ($n=1000$, $t=1\dots 10^4)$ of population composition for different protester types and their mean DIME variables for two parameter regimes. Left panels (a, c, d): High movement success, high individual + collective re-framing, $p=0.2, F=0.2, \phi = 0.2$. Right panels (b, e, f): High movement failure, low individual + collective re-framing, $p=0.8, F=0.8, \phi = 0.8$. In (c), we see a diverse range of protester types, with latent conventional (LaCo), active innovators (AcIn), and active conventional (AcCo) being most prevalent. In (e), the population is comprised almost entirely of latent radical protesters, i.e. they are inactive but orientated to radical tactics if they were to become active.
• Figure 3: Mapping the emergence of the six protester types under varying probability of failure signals ($p$) and threshold for individual re-framing ($F$). Contour plots show population composition (numbered lines, colourbars) at the long-run steady state for the six different agent types under study, aggregating simulations at a given ($p,F$) location ($n=1000$): active conventional (a), active innovators (b), active radical (c), latent conventional (d), latent innovators (e) and latent radicals (f). Collective re-framing is held constant across plots ($\phi = 0.8, R = 10$). The red square marker indicates the intransigent authority context studied in the right panels of Fig. \ref{['fig:NetworkTimeSeriesComposite']} ($p=0.8, F=0.8, \phi = 0.8, R = 10$). Contours show an interaction between probability of failure signals ($p$) and threshold for individual re-framing ($F$).
• Figure 4: Dominant types of agent under varying probability of failure signals ($p$) and threshold for individual re-framing ($F$). (a): Variation in dominant action, utilising colour and hatching (see Key) for the six agent types, with $n=1000$ and varying probability of failure signals ($p$) and threshold for individual re-framing ($F$) over [0,1] (holding collective re-framing at $\phi = 0.8$ and $R = 10$). (b): Population composition (on z-axis) for the corresponding dominant strategy at each parameter pair. The latent radical dominant region shows the highest population fraction of dominant action of any regime.
• Figure 5: The role of collective re-framing in the emergence of dominant protest types. Each panel depicts the population composition of the dominant action in the steady state, for a given probability of failure signal---threshold for individual re-framing pair ($p,F$) ($n=1000$). Rows and columns show the outcome when varying: (left to right) the threshold for collective re-framing ($\phi \in \{0.2, 0.5, 0.8\}$); and (bottom to top) the length of free communication ($R \in \{1, 2, 10\}$). Colouring follows the key given, covering the six agent types under study. Overall, we can see the largely resilient nature of the dominant latent radical region, under variation in collective re-framing.