The Influence of Semantic Structure on Generalization

Authors: Brooklyn Stisser, Sonya Zin, Sarah Solomon

Field of Study: Science, Technology, Engineering, and/or Math

Faculty Mentors: Sarah Solomon

Easel: 9

Timeslot: Afternoon

Abstract: Semantic categories (e.g., beetle, butterfly) can have complex internal structures, but how this structure influences semantic cognition more broadly is still unclear. Two novel “insect” categories were designed with two different structures (i.e., Modular, Lattice) to test the theory that category structure influences the generalizability or “stretchiness” of that category. The categories were defined by discrete visual features (e.g., wings, antennae) and the cooccurrences among those features determined the category’s structure. Participants learned the two categories in a missing feature task and then were asked to categorize novel exemplars that contained combinations of features from the two learned categories. Importantly, some exemplars have an equal number of features from each category, making it possible to test whether category structure pushes generalization in one direction or another. Examining how people categorize these hybrid exemplars will reveal whether category structure can make some categories more generalizable than others.