Statistical learning and chunking shape language and reading across development
How do individuals acquire language in the face of its perceptual challenges? In this talk, I argue for the contribution of chunking—a basic memory process—to the statistical learning of language across the lifespan. My research illustrates how statistical learning and chunking work together to enable the acquisition of a variety of linguistic structures, and how proficiency in statistically-facilitated chunking predicts individual differences in speech processing. Such computations further extend to the learning of written regularities in both children and adults, and even predict performance on established measures of reading. Together, these studies highlight the fundamental contribution of basic, domain general computations to sophisticated linguistic abilities across modalities and development—both in the lab and in the real world.