SCEW Postdoctoral Fellow
Lewis (1969) introduced signaling games to show how communication emerges when rational agents imbue simple actions with conventional meaning. Since then, signaling games have given rise to a rich field of study. In this talk, I first provide a brief introduction to signaling games. I then propose versions of the game that help us understand some large-scale patterns observed in natural languages—namely, Zipf’s law relating word frequency and rank, and the negative correlation between word frequency and rate of change. To conclude, I single out open questions and sketch tentative ways to test the model framework experimentally.