Sharp transition towards shared vocabularies in multi-agent systems

Abstract

What processes can explain how very large populations are able to converge on the use of aparticular word or grammatical construction without global coordination? Answering thisquestion helps to understand why new language constructs usually propagate along anS-shaped curve with a rather sudden transition towards global agreement. It also helps toanalyse and design new technologies that support or orchestrate self-organizingcommunication systems, such as recent social tagging systems for the web. The articleintroduces and studies a microscopic model of communicating autonomous agentsperforming language games without any central control. We show that the systemundergoes a disorder/order transition, going through a sharp symmetry breaking process toreach a shared set of conventions. Before the transition, the system builds upnon-trivial scale-invariant correlations, for instance in the distribution of competingsynonyms, which display a Zipf-like law. These correlations make the system readyfor the transition towards shared conventions, which, observed on the timescaleof collective behaviours, becomes sharper and sharper with system size. Thissurprising result not only explains why human language can scale up to verylarge populations but also suggests ways to optimize artificial semiotic dynamics.