Research

My research explores different facets of language representation and comprehension. Some questions I'm interested in include: How do neural systems learn and represent content and structure? How do different knowledge types flexibly and quickly assembled into new configurations to help us with a wide range of tasks, like solving problems, or understanding a work of fiction?

Neural bases of language comprehension

I explore the neural bases of syntactic and semantic representations as a window into our generative conceptual capacity. As it is an interdisciplinary endeavour, I try my best to do my homework, grounding my work in linguistic theory, attending to constraints from neuropsychology, and borrowing designs and paradigms from cognitive science and psycholinguistics. [Papers: Law & Pylkkänen, 2021; Tulling, Law et al., 2021; Law et al., 2026]

Structure learning in neural systems

More recently, I began seeking explanations in computational terms, using tools from deep learning to test intuitions about our semantic competence and to better characterise its computational bases. I'm currently working on a project exploring how neural systems learn structural representations like number and magnitude.

AI and human and animal cognitive science

My colleagues and I have also recently argued that AI can benefit from learning from human and animal cognition research, and we outlined a way forward in a comparative approach to the evaluation of large language models, presented in the NeurIPS 2024 Behavioral ML Workshop and published as an ICML 2025 archival paper. See also Melanie Mitchell's Substack discussing our work.