Unified opinion formation analysis in rewriting logic

Processes of opinion formation rooted in social dynamics can significantly contribute to the polarization of social, political, and democratic interaction. Opinion dynamic models are essential for understanding the impact of specific social factors on the acceptance or rejection of opinions. This extended paper builds upon the conference presentation documented in [1], introducing improvements and new opinion models that explore biases and collective human behaviors. It presents a framework based on concurrent set relations that formalizes, simulates, and analyzes social interaction systems with dynamic opinion models. Within this framework, standard models for social learning are realized as specific instances. Implemented in the Maude system as a fully executable rewrite theory, the framework enables a detailed examination of how agents' opinions can be influenced within a system. The authors report on new formalization of several and existing social learning models, exploring their relationships with different concurrency models. New experimentation involving reachability analysis, probabilistic simulation, and statistical model checking has been conducted. These experiments are crucial for validating significant properties related to dynamic opinion models in Maude, offering new insights into the mechanisms of opinion shaping in social interaction.