The rational analysis of categorization and the ACT-R architecture
John R. Anderson, Michael Matessa · 2000
Abstract As the cognitive revolution psychologists have striven to develop detailed theories of the information-processing steps involved in performing a particular task. The goal is to be able to say moment by moment what is happening in the minds of people as they do the task. While this goal is laudatory there are serious problems with how feasible it is. When one starts to deal with complex tasks (like doing high school mathematics-Anderson et al., 1995) the sheer volume of the detail can be overwhelming. Even in simpler laboratory tasks, the data that one collects is often at a much coarser grain size than the processes one is postulating. There arise identifiability problems about whether very different ways of processing information might not result in the same data. Individual differences are another problem. When one gets to this level of detail it is unlikely that every subject is doing the task in the same way.