Our projects aim to understand the early emergence of human knowledge by focusing learning processes in the context of children's word learning and language learning. Children begin word learning by learning each word slowly and needing many repetitions and instances to know the range of things in the lexical category; but they soon become rapid, one-trial, learners of different types of words; nouns, verbs, and adjectives. Our work suggests that this teaches the child how to learn new words. Our recent collaborative work is to tie this idea to the possible creation of self-sustaining learning system-robotics. With this project, we would further study how the learning system emerges through the regularities in the language learning environment, by comparing those regularities across different languages, by studying how words (and other cues) guide attention in the learning moment, and through experimental and training studies with children and adults.
As for our most recent project, the work concentrates on a fine-grained analysis of the relevant information in the learning environment in different language learners. The ongoing studies seek to move beyond merely words and their referents and syntactic contexts to include the dynamics -the temporal relations among hearing names, looking at, and acting on objects. Ideally, we need to specify this learning environment from the child's point of view. Accordingly, as a part of the project, we are completing the technique of putting a tiny camera on the forehead of children and infants (calibrating it to encompass the visual field of the child) and recording the child's interactions with the parent in object play. We hope to document how the stream of words (nouns and verbs) lines up with attention to objects and actions in the two languages and the two learning environments.