Representing a complex environment, with abundant interdependencies and subtle regularities requires a rich set of categories. Many categories have members that do not seem to fit the category in one way or another. The category of mammals, for example, includes familiar members such as lions and tigers and bears, but also includes unusual members such as bats and whales. The proposed research will improve our understanding of how people learn and represent these kinds of overlapping categories and their rogue members. Experiments address: the impact of varying the number of normal members and the number of rogues, the importance of rogues being perceptually distinct, the impact of the locus of category and item information within the stimulus complex and the affect of different training orders on learning ease. The latest mathematical theories of categorization will be used to model the results. This research can provide valuable insight into learning and representation in overlapping categories and so is important for understanding basic principles of learning that extend through all areas of higher cognition, such as face recognition, language development and concept learning. Understanding how these cognitive systems function will improve our ability to assess and treat cognitive disorders. In addition, this research can help to inform training procedures for educators and medical health professionals performing diagnoses, and therefore may provide a positive public health benefit from two independent directions.
