Language and friendships: A co-evolution model of social and linguistic conventions

Abstract

Human social networks are dynamic, yet the majority of computational models of language evolution operate with either fixed or random interaction patterns between agents. This paper presents a model of lexicon formation in which agents can adjust their social ties based on the success of performed interactions. Simulation results show that different levels of `friendship? between agents co-evolve with a system of linguistic conventions, with the observed social network formation processes largely resembling those of modern humans.