How Children’s Variable Input Shapes Phonological Structure: Language Learning Where It Happens
Children learn the patterns of their native language(s) through years spent interacting and observing in their everyday environments. To understand how children build phonological systems, we must then study language learning where it happens: in homes, communities, and diverse linguistic environments. This talk investigates how children’s variable input shapes phonological development using evidence from large-scale recordings of natural language environments. First, drawing on a massive cross-linguistic database of infant babbling, I show how early vocal development follows similar trajectories despite dramatically different linguistic inputs. Next, examining bilingual Quechua-Spanish children in the Bolivian highlands, I demonstrate how differences in early language exposure affect children’s phonological working memory and processing abilities. Finally, I present new work examining how infants who are deaf and receive cochlear implants develop speech in their everyday environments, using speech synthesis and signal processing techniques to reverse-engineer early language input and understand how learning unfolds despite degraded sensory conditions. Throughout, I highlight innovative computational tools I’ve developed for collecting and analyzing naturalistic speech data outside traditional lab settings—methods that advance both developmental research and linguistic fieldwork more broadly. Overall I emphasize how studying language acquisition in diverse, real-world contexts reveals how children can build phonological systems from variable input.