The evolution of efficient compression in signaling games

Converging evidence suggests that natural language meaning systems are efficient by jointly maximizing cognitive simplicity and communicative informativeness. Comparatively less is known about how languages might optimize over time for communicative efficiency. Our goal in this paper is to use minimal dynamic models to give a high-level description of the evolution of efficient meaning systems. To do this, we provide a model of emergent communication combining evolutionary game theory with a recent information theoretic account of efficiency in semantic systems. We perform simulations of adaptive dynamics requiring minimal assumptions about agents’ cognitive resources, and observe that emergent languages converge near the achievable bounds of efficient compression. This unifies existing accounts of communicative efficiency with minimalist accounts of how meaning can emerge ex nihilo.