Fragmentation transitions in multi-state voter models

Adaptive models of opinion formation among humans can display a fragmentation transition, where a social network breaks into disconnected components. Here, we investigate this transition in a class of models with arbitrary number of opinions. In contrast to previous work we do not assume that opinions are equidistant or arranged on a one-dimensional conceptual axis. Our investigation reveals detailed analytical results on fragmentations in a three-opinion model, which are confirmed by agent-based simulations. Furthermore, we show that in certain models the number of opinions can be reduced without affecting the fragmentation points.

💡 Research Summary

The paper extends the adaptive voter model—a paradigmatic framework for opinion dynamics on co‑evolving networks—to the case of an arbitrary number of discrete opinions, allowing the interaction strength (rewiring probability) to depend on the specific pair of opinions involved. In the classical two‑state adaptive voter model, a single rewiring rate p controls the relative speed of link updates versus opinion updates, leading to either a globally homogeneous consensus or a fragmented state composed of two internally homogeneous components. The authors point out that real‑world issues often involve more than two alternatives and that the “distance” between opinions can be heterogeneous; therefore they introduce a set of pairwise rewiring probabilities p_{ij}=p_{ji} for every unordered pair (i,j) of opinions. A high p_{ij} indicates a strongly antagonistic pair (links are likely to be cut), while a low p_{ij} reflects similarity (links are more likely to be retained and opinions may be adopted).

The model is defined on a random graph with N nodes and L undirected links (mean degree k=2L/N). Each node α carries an opinion s_α∈Γ={g₁,…,g_G}. At each elementary step a random link (α,β) is selected. If the two endpoints share the same opinion the link is inert; otherwise it is active and either (i) the link is rewired with probability p_{ij} (where i and j label the opinions of α and β) or (ii) one endpoint adopts the other’s opinion with probability 1−p_{ij}. Rewiring creates a new link from the focal node to a randomly chosen node that already holds the same opinion, preserving the symmetry of the opinion adoption process.

The authors focus first on the smallest non‑trivial case G=3 (opinions A, B, C) with three distinct rewiring rates p₁=p_{AB}, p₂=p_{AC}, p₃=p_{BC}. They identify five possible asymptotic outcomes: (i) a fully active regime that eventually reaches global consensus, (ii) full fragmentation into three homogeneous components, and (iii) three variants of partial fragmentation where one opinion forms an isolated homogeneous component while the remaining two opinions continue to interact. To analyse the stability of these states they adopt the “active‑motif” approach introduced in earlier work (e.g., Ref.