Language adapts to pressures from production: experimental and computational evidence

Abstract

Languages have the daunting task of conveying an infinity of possible meanings. Yet at the same time, they are constrained by the limits of human memory, perception, and motor control. In this thesis, I study how language structure emerges from a complex interplay between these constraints. I argue that language is shaped by pressures for ease of retrieval and articulation, pressures which stem from the cognitive demands of real-time language production.

In three projects, I study how the challenges associated with language production shape the way we learn languages, the way we use them in real-time communication, and the way they change and evolve over time. I do this in two main ways. The first involves asking human participants to learn miniature made-up languages, and then seeing what happens when they have to produce these languages themselves. The second involves using computer code to simulate the process of human communication, to see how the effects I observe in my experiments might accumulate over thousands of generations.

Through this combination of empirical methods, I show that, both at an individual-level and at a population-level, language structure is fundamentally a balancing act between competing pressures.