Extracting prototypes from lexical feature norms for settlement concepts.

The present study explored whether people share a common understanding of different settlement concepts despite individual variation. Participants completed a property listing task where they were asked to generate features for 57 settlement concepts (e.g., "city," "college town"). We used hierarchical cluster analysis to identify distinct clusters based on shared features. Central tendencies extracted from the clusters at different levels of abstraction revealed featural prototypes and an overall family resemblance structure. To probe the effects of regional context on conceptual structure, we used a subset of participants who were long-term residents of Canada or the USA. We found that prototypical features varied regionally, suggesting an effect of participants' geographical region on conceptual structure. This work is the first to investigate how people represent settlements as concepts and helps centralize the utility of semantic feature norms in understanding how people collectively think about where they live, and the importance of context effects on representations of settlements.