Dual language input, semantic structure and word learning in typically developing and late talking bilingual children

Project Summary Over the last several decades, developmental language scientists have sought to understand the effects of language input on children’s vocabulary outcomes. Much of this work has been correlational and has focused on static standardized measures of vocabulary size in typically developing children, often excluding children with smaller vocabularies, referred to as late talkers. Although vocabulary size is an important predictor of later outcomes, it provides limited insight into children’s semantic structure—that is, how word meanings are represented, organized, and connected in the mind. Computational and empirical studies suggest that children’s semantic structure (1) reflects the statistical regularities and semantic relationships in their language environments, and (2) predicts word learning beyond what is explained by vocabulary size alone. However, it remains unclear how children’s lexicons reflect language input and whether different semantic structures support distinct patterns of word learning. The primary objective of this proposal is to examine the interactions between language input, semantic structure, and word learning in toddlers who are typically developing and late talking. We will test 80 toddlers between 24 and 30 months of age, equally divided into typically developing and late-talking groups. In Aim 1, we will use semantic network approaches to model children’s lexicons and characterize the relationship between language input and semantic structure. In Aim 2, we will examine how semantic structure relates to statistical word learning. We will analyze whether children’s semantic network properties predict performance on word learning tasks. In Aim 3, we will assess whether the relationships among language input, semantic structure, and word learning differ between typically developing children and late talkers. This career development proposal includes a multidisciplinary team of mentors across Psychology, Computer Science, Communication Sciences and Disorders, Education, and Public Health. The candidate will receive training in observational methods for studying parent-child interaction, experimental and eye-tracking methodology for toddlers, and network science approaches. The long-term goals of this work are to track language development longitudinally in children and to inform development of novel interventions using behavioral and network-based methods. These goals align with NIDCD’s strategic priorities, including enhancing our understanding of typical and disordered communication processes (Theme 1) and advancing data science methodologies in communication research (Theme 5).