Unifying Local Dynamics in Two-State Spin Systems

Abstract

We present a new two-state {+-} opinion dynamics model which defines a general frame to include all local dynamics in two-state spin systems. Agents evolve by probabilistic local rules. In each update, groups of various sizes k are formed according to some probability distribution {a_k}. Given a specific group with an initial (j) agents sharing opinion {+} and (k-j) agents opinion {-}, all k members adopt opinion {+} with a probability m_{k,j} and opinion {-} with (1-m_{k,j}). A very rich and new spectrum of dynamics is obtained. The final opinion is a polarization along the initial majority, along the initial minority or a perfect consensus with an equality of opinions as function of the parameters {a_k, m_{k,j}}. In last case, two regimes exist, monotonic and dampened oscillatory. The transition from polarization to consensus dynamics occurs for values of the parameters which reproduce exactly the Voter model. A scheme is presented to express any local update in terms of a specific set {a_k, m_{k,j}}. Most existing opinion models are exhibited at particular limits.