Stochastic Naming Game and Self-constraining Nature of Synonymy

Abstract

Using language games has been a fruitful lane of research to investigate the social aspects of language and communication. Our approach aims to start covering the ground between the elementary naming game and the complex models for social use and evolution of linguistic structure such as FCG. It uses weighted lists as lexicons, probabilistic choice as a selection mechanism and lateral inhibition as the weight update scheme. The results demonstrate that the model is a generalization of the elementary naming game and provide a clearer picture of how big a lexicon agents use for the task, how this size can be controlled using the model parameters and a possible way of explaining how synonymy may be controlled.


Back to Table of Contents