The Knowledge Effects In Category Formation Of Children

Categorization and category formation is the commencement of people's thinking and the essential mode to store and integrate the knowledge of objects systematically. By the influencing of the similarity-based view, most of prior researches on category formation focus on empiric-driven (bottom-up) processing. But after the theory-based view was suggested, theory-driven (top-down) processing in category formation is obtained much more attention from 1990s. Some researches show that category structure or category representation is affected profoundly by prior knowledge and naive theory. But we still do not understand the role of prior knowledge in children's category presentation clearly, as well as the mechanism and the development of knowledge effects during the process of children's category formation.To figure out the issues mentioned above, 3 multi-factor experiments are designed to examine the knowledge effects in children's category formation as well as the mechanism and development of knowledge-based processing. Each experiment uses features-theme paradigm. And the artificial categories are family resemblance categories. Subjects include 3 age groups from 4, 6 and 8 grade respectively of primary school and middle school. Experiment 1 examined the knowledge effects in category construction under unsupervised learning condition. Experiment 2 examined the knowledge effects in category learning under supervised learning condition. And experiment 3 examined the role of knowledge-related features and rote features in category learning. What the results suggested are as follows:(1) Prior knowledge plays an important role during category formation of children.(2) Children would rather construct category by themes than by appearance similarity if parts of the category features can be related to one theme on the basis of prior knowledge. And more the knowledge-related features exist in category, more the children categorize by themes.(3) Children's performance of category learning is advanced apparently if the theme shared by the knowledge-related features is consistent with the category structure. And more the knowledge-related features exist in category, more the progress is made by the children.(4) The development of knowledge-based processing in unsupervised category learning is different from supervised one. Namely, the subjects are observed the same amount of knowledge effects in category construction among different age groups, but the knowledge effects in category learning are different.(5) Not only the information has been acquired on the knowledge-related features but also it does on the rote features. So it suggests that children have learned category structure by the approach of knowledge first and features integrated.(6) The knowledge effects of category learning are influenced by children's knowledge base, as well as by children's capability of information encoding and retrieving.(7) Domain specificity of artificial category has no effect on children's category construction and learning.In general discussion we give some suggestions on the role of prior knowledge and theories in category coherence, the category representation of children, the age difference of knowledge-based processing, the integration of prior knowledge and empirical information, and some consideration on method in this field.