The associated project will study the dynamics of linguistic conventions in scenarios where agents interact that use different signaling strategies or have different preferences regarding the extra-linguistic contextual setting. Game theory, especially the theory of signaling games, will be used as the primary conceptual framework to carry out these investigations. The main part of the work program of the project concerns applications of evolutionary game theory to explain the emergence of linguistic structures in a population of interacting agents. This framework is applicable in situations where (a) the agents within the population use different strategies, but (b) communication is nevertheless partially successful, to a varying degree. Communicative success is identified with fitness in the Darwinian sense. In other words, the degree of communicative success determines the likelihood with which a certain linguistic feature is imitated later. This creates a positive feedback loop that eventually leads to convergence of a common convention. The guiding hypothesis of the project is that the targets of a such an (cultural) evolutionary dynamics correspond to grammatical features that are typologically common among the languages of the world. On the modeling side, the project will work with analytical models and with agent-based computer simulations. Empirically we will draw on quantitative data from cross-lingusitic and diachronic studies.
For further information, please contact Gerhard Jäger (project leader): gerhard dot jaeger at uni minus tuebingen dot de