A Rate-Distortion view of human pragmatic reasoning

TL;DR

This paper analyzes the Rational Speech Act framework for pragmatic reasoning, revealing its connection to optimization principles and Rate-Distortion theory, thereby providing a mathematical foundation for understanding human pragmatic behavior.

Contribution

It demonstrates that RSA recursion implements an optimization of utility versus effort and grounds RSA in Rate-Distortion theory, advancing the theoretical understanding of pragmatic reasoning.

Findings

RSA recursion performs alternating maximization between utility and effort

Expected utility does not always improve with recursion depth

RSA can be grounded in Rate-Distortion theory, aligning with human behavior

Abstract

What computational principles underlie human pragmatic reasoning? A prominent approach to pragmatics is the Rational Speech Act (RSA) framework, which formulates pragmatic reasoning as probabilistic speakers and listeners recursively reasoning about each other. While RSA enjoys broad empirical support, it is not yet clear whether the dynamics of such recursive reasoning may be governed by a general optimization principle. Here, we present a novel analysis of the RSA framework that addresses this question. First, we show that RSA recursion implements an alternating maximization for optimizing a tradeoff between expected utility and communicative effort. On that basis, we study the dynamics of RSA recursion and disconfirm the conjecture that expected utility is guaranteed to improve with recursion depth. Second, we show that RSA can be grounded in Rate-Distortion theory, while maintaining a similar ability to account for human behavior and avoiding a bias of RSA toward random utterance production. This work furthers the mathematical understanding of RSA models, and suggests that general information-theoretic principles may give rise to human pragmatic reasoning.