Doctoral Dissertation Research: Using computational modeling for investigating the acquisition of Wh-dependencies in sentence structures

This doctoral dissertation project focuses on how children learn “wh-dependencies,” a challenging aspect of language that can be found in questions like “What is she reading?” Understanding these structures, especially in long questions such as “What does her mother think she is reading?” requires children to grasp intricate relationships between words and hold information in memory while comprehending longer sentences. Remarkably, by age four, children achieve this feat despite two major challenges: (1) cognitive limitations, such as reduced memory capacity compared to adults, and (2) variations in language input, as children are exposed to widely different linguistic environments but still succeed in learning these complex relationships. This research investigates theories of language learning and examines whether they can account for children’s ability to overcome these challenges. The findings provide valuable insights into the processes of language acquisition and child development, as well as practical strategies to support children’s learning. The project employs computational cognitive modeling to implement a learning theory of wh-dependency acquisition. This approach concretely specifies how the learner processes wh-dependencies in child-directed speech and how the learner outputs behavior that demonstrates adult-like wh-dependency knowledge. To implement the two sources of limitations outlined above, this research tests the performance of the learning model under child-like memory limitations and examines how the model behaves with input across different environmental settings. To ensure the learning theories reflect real-world scenarios, the research also expands a database of child-directed speech to better capture the input children encounter. By precisely implementing learning theories and aligning them with limitations children face, this project aspires to uncover foundational principles of language learning and further our understanding of the cognitive mechanisms that support this remarkable ability. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.