Abstract
Children acquire a great deal of their knowledge from the input of others, and they do so in a selective way. How do children decide from whom to gain knowledge? The aim of this thesis is to provide the groundwork for an integrative account of a dynamic process underlying children’s selective learning. Across a series of three studies, I investigated 3- to 5-year-old preschoolers’ evaluations of human beings and humanoid robots as informants for learning purposes. The first study provides evidence that having prior knowledge about the information to be learned affects children’s subsequent willingness to accept information from a questionable source, suggesting the vital role of self in selective learning (Chapter 2). The second study shows that children adjust their learning strategies depending on the character of their informant (e.g., a person vs. a robot); they are less willing to learn from a robot that previously made naming errors than a person who made the same errors (Chapter 3). Finally, the third study demonstrates that children can infer the knowledgeability of robots by monitoring the robots’ learning processes, and they subsequently use such inferences about the robots’ knowledgeability to guide their own learning from the same robots (Chapter 4). These studies demonstrate how young learners make use of data from themselves, others, and interactions with other social beings in ways that affect their evaluations of informants and their selective learning.