Skip to main content Skip to secondary navigation
Main content start

David Rose

One of the most influential ideas about concepts is that you cannot characterize them by appealing to observable features alone. Instead, concepts encode essential properties or properties concerning what makes something the kind of thing that it is. My view is that thinking about what things are for plays a fundamental role in categorization and that essences are associated with purposes. If that is right, then this has a broad range of consequences for the way philosophers, psychologists, and computer scientists think about the nature of concepts and categorization. My research is aimed at tracing out the developmental roots of essentialist thinking and the ways in which it changes—and stays the same—throughout adulthood, as well as whether state-of-the-art-language models, like Generative Pre-trained Transformer 3 (GPT3), show signs of engaging in essentialist categorization. If so, what does it mean for how we should think about the nature of concepts and categorization?