How is our knowledge organized, and how does it develop?  This talk will explore
these questions through the study of dissociations in children's behavior. 
Children can appear very smart when tested with certain tasks, while appearing
completely oblivious when tested
with other tasks meant to measure the same knowledge.  Such dissociations cut
across a range of domains (e.g., memory, rule use, and spatial processing),
measures (e.g., looking, reaching, navigation, verbal response), and ages (being
particularly salient from infancy through childhood).  I will contrast two
approaches to such dissociations.  The first approach focuses on the role of
deficits in modules or systems that are isolated from intact knowledge; the second
approach focuses on the role of graded knowledge representations.  I will present
research designed to evaluate these approaches, and discuss implications regarding
the organization and development of knowledge.