Critical Slowing Down, Not Intensity, of Collective Anger Predicts Surges of Social Movement Activity

<title>Abstract</title> <p>Social movements often exhibit abrupt surges in activity that are difficult to anticipate. Drawing on complex systems theory, we examined whether such surges can be predicted from the temporal dynamics of collective emotions. Using approximately 60 million #BlackLivesMatter tweets collected between 2015 and 2020, we compared two potential predictors of activity surges: anger intensity and anger persistence. Anger persistence was indexed using critical slowing down indicators, including autocorrelation and variance. Mean levels of anger did not predict surges. In contrast, increasing autocorrelation in collective anger reliably preceded surges in activity, indicating that emotional persistence rather than emotional intensity provides predictive information about impending mobilization. Multiple robustness analyses showed that this effect was not attributable to preprocessing artifacts. These findings extend early warning signal research to large-scale human behavior and suggest that collective emotional dynamics can provide advance warning of transitions in social systems.</p>