Identifying cross-situational word learning mechanisms across development

点击翻译按钮获取中文摘要

Project Summary/Abstract Vocabulary acquisition is a crucial component of children’s early development, laying the groundwork for early success in school and enabling children to communicate with and learn from others. While children rely on many word learning strategies to build their vocabularies, recent research has revealed a powerful role for cross-situational word learning. Even when a word’s meaning is ambiguous in one context (e.g., a child hears “It’s a dog!” and several animals are present), children can use a series of these ambiguous utterances to learn the word. However, the mechanism by which children do so is not clear: the Cumulative Statistics theory proposes children track word-meaning co-occurrences across contexts while the Hypothesis Testing theory proposes children collect evidence for or against only one hypothesized meaning, or previous hypotheses. The proposed studies use behavioral, eye-tracking, and computational modeling methods to evaluate these rival word learning mechanisms. The studies take a developmental approach, asking how children’s word learning abilities change from 2 to 7 years of age and into adulthood, as well as what learning mechanisms best account for these changes. This work will also provide the first evidence regarding whether children can learn verbs across semantically ambiguous contexts: novel verbs can refer to separate aspects of a single event (such as the manner or path of motion), presenting a challenge for children’s learning. Finally, we examine a related prediction from the memory literature: that word learning will be improved when learners encode a word’s potential meanings in an integrated event representation. These studies address fundamental questions about cross-situational learning’s mechanisms, scope, and underlying representations throughout development. These findings will offer insight into how children learn words during an important period of language development, stretching from preschool to 2nd grade, as well as how these abilities mature in adulthood. By examining both nouns and verbs, these studies also provide a crucial test of how word learning strategies previously documented for nouns extend to children’s acquisition of verbs, which pose additional challenges and, as such, are later-acquired and can be especially hard for children experiencing language delays to learn. In addition, these studies will reveal how the structure of word learning contexts, which often involve potential word meanings being involved in a shared event, can enhance both noun and verb learning. Finally, these studies will provide the applicant with critical postdoctoral training to advance his career. The applicant will master the new methodological and theoretical skills necessary to examine language learning throughout the lifespan and construct computational models of major language learning theories. This training will be highly interdisciplinary, utilizing mentors, coursework, and professional development activities across multiple departments and research organizations. This training will ultimately prepare the applicant to become a successful independent researcher.