Simulation of the Emergence of Language Groups Using the Iterated Learning Model on Social Networks

In evolutionary linguistics, the Iterated Learning Model (ILM) is often used for simulating the first language acquisition. Our purpose in this paper is to develop an agent-based model for language contact based on ILM. We put a learning agent on each node in the social network. Our experimental result showed that the language exposure rather deteriorates the emergence of local common languages, and grammars become non-compositional, which is different from our expectation. However, we have shown that an excessive string-clipping as well as a language exposure may constrain the appearance of local language community, independent of the shape of networks. Keywords–Simulation, Language Acquisition, Iterated Learning Model, Social Network.