The spatial distribution of clusters and the formation of mixed languages in bilingual competition

We use cellular automata simulation methods to study the competition between two languages (language A$A$ and B$B$). We assume each of the two languages consists of F$F$ independent features and define an individual as two F$F$-length “identity level” integer strings. The value of each integer of the strings indicates whether the individual is willing or unwilling to express a certain feature and how his willingness or unwillingness is. In our model, individuals who speak either language A$A$ or B$B$ are randomly placed on a square lattice initially and programmed to evolve under the communication and interaction methods. First, we consider the situation that the competition occurs between two languages with only one feature. We find the differences between short-time coexistence processes and long-time coexistence processes and discuss how the spatial distribution of languages evolves. We observe that periodic line-shaped boundaries take shape in some simulations and lead to long-time coexistence between the two languages. Then we study the multi-feature cases when F=10$F=10$ and find the correlation among the evolution of different features decreases with time. We also observe the extinction of two primitive languages and formation of mixed languages.