Computational linguistics analysis of leaders during crises in authoritarian regimes


We investigated linguistic patterns in the discourse of three prominent autocratic leaders whose tenure lasted for multiple decades. The texts of Fidel Castro, Zedong Mao, and Hosni Mubarak were analyzed using a computational linguistic tool (Coh-Metrix) to explore persuasive linguistic features during social disequilibrium and stability. The analyses were guided by the elaboration likelihood model of persuasion, which contrasts central versus peripheral routes to persuasion. Results show these leaders utilize the central persuasion route, with more formal discourse patterns during times of crises versus non-crises. A significant interaction between leader age and armed conflict revealed interesting adaptive characteristics. Specifically, leaders' formality decreases over time in both crises and non-crises times, but this attenuation is less prominent during crisis periods. The implications of these results are discussed in the context of using computational linguistics analyses to generate potential predictive models of social disequilibrium and to advance our understanding of authoritarian regimes.

Publication Title

Dynamics of Asymmetric Conflict: Pathways toward Terrorism and Genocide