I am interested in the acquisition and representation of conceptual knowledge: our general knowledge about categories of things and their properties. In the brain, conceptual knowledge is not a single, homogeneous system: it is instead organized into components responsible for specific domains (living; non-living) and attributes (color; shape). One line of my research asks about the principles that could explain such divisions and their cortical placement; in particular, by focusing on non-concrete attributes such as the purposes of objects, the goals of mentalistic actions, and belief traits of people. Relative to sensory-related knowledge, the large-scale organizing factors of such properties are much less understood, and finding them could reveal general principles of how higher-level cognition is neurally implemented.

My more recent work investigates the acquisition of non-concrete attributes. How do we learn that telephones enable communication, that people have personalities, or that plants need sunlight to grow? Such knowledge has to be learned from experience, but is not given directly by sensory qualities. One hypothesis is that it relies on inferences over the statistical structure of events surrounding such objects. For example, communication could refer to a reliable contingency between speaking and receiving a reply. But we must identify such structure in a noisy environment with many uninformative events, and draw generalizations of this structure across different specific experiences. My current work looks at the neural mechanisms that allow us to go from seeing contingencies among specific events, to those that generalize those relationships across radically different events.