Language Models Generalize to Human-like Word Order Preferences

Amanda Popadich, Shane Steinert-Threlkeld

A central question in language acquisition is whether linguistic biases can emerge from general learning mechanisms operating over underdetermined input. Artificial Language Learni…