Multi-dimensional multi-option opinion dynamics leads to the emergence of clusters in social networks

In real-world social networks, opinions evolve within a multidimensional space of multiple topics being concurrently discussed, and a multi-option decision-making process, rather than a simple binary choice, takes place. Our work introduces a multi-dimensional multi-option opinion dynamics model capturing the complexity of opinion evolution in social networks. The model exploits the coupling of inner opinion and outward action, emphasizing how similar actions strengthen interactions between agents. Unlike existing research, in which consensus, clustering or polarization result from specific network structures, we find that different attitude patterns towards neighbours lead to the spontaneous emergence of such macroscopic phenomena, which are therefore independent of network structural features. We provide analytical conditions for the transitions to these behaviours, confirming them via simulations on different networks. Thus, our model allows one to explain the emergence of collective phenomena observed in real-world situations, thereby providing insights in areas such as opinion guidance and multi-agent decision-making. • First opinion model with realistic dynamics of beliefs, actions and choices. • Mathematical explanation of how the model reproduces real-world phenomena • Independence of the results from the structure supporting the social interactions.