This project is concerned with how people learn about, use, and update their concepts and categories. Categorization is fundamental to reasoning, because human language and thought is made up of and organized around concepts. A particular focus of the project is to understand more fully how previous knowledge and experience are integrated with new information to determine concepts. For example, much of science education involves learning new terms and concepts, and there is considerable evidence that previous knowledge sometimes leads to misconceptions and misinterpretations of new information. In other cases, prior knowledge is unambiguously helpful. Clearly we need to know much more about how knowledge and experience combine to determine what is learned. The research project has several distinct aims. One goal is to evaluate what makes some categories easy and natural to learn and what makes other categories unnatural and difficult to learn. A related goal is to develop further and to test quantitative and computational models of category learning. One set of studies will evaluate these models in contexts where the learning examples have a complex, hierarchical structure. A second line of work will study the effects of different perspectives on the perceived depth or centrality of different properties of concepts. A third set of experiments will study how abstract advice or theories is combined with concrete examples in category learning. The category descriptions associated with theories (or diagnostic criteria in psychodiagnostic classification) are often much more abstract than the properties or behaviors of individual instances of a category. The overall goal of the project is to understand how these abstract knowledge sturctures become linked with specific, contextualized experience to determine concept formation. Potential applications of this work range from science education to any of a number of situations where diagnostic classification is an important activity.
