Statistical Learning in Language Acquisition

点击翻译按钮获取中文摘要

DESCRIPTION (provided by applicant): Acquiring a language is among the most daunting feats uniformly accomplished by our species. Explorations of the language learning process present the opportunity to address central issues pertaining to human cognition and its development. The proposed research represents one approach to this topic: the study of the architecture of the learning processes underlying language acquisition, using laboratory learning experiments performed with adult, child, and infant participants. The goal of this research program is to substantially increase our knowledge of the statistical learning mechanisms that detect linguistic units by tracking the patterns of sounds, words, and other units in the input. Recent results suggest that statistical learning processes play an important role in the acquisition of language. However, little is currently known about the types of statistical regularities computed by learners and the constraints on learning that support successful knowledge acquisition. The proposed experiments will ask: (1) How does the structure of the input constrain statistical learning? Studies will consider perceptual, linguistic, and informational constraints on the choice of cues as input to learning. (2) How does prior experience influence statistical learning? Manipulations of prior experience, inside and outside the lab, will be used to assess effects on subsequent learning. (3) Can statistical learning account for basic phenomena in language acquisition? Interactions between statistical learning and other types of processes, such as rule learning and the presence of other cues in the input, will be investigated. All of these issues will be addressed using previously developed laboratory learning paradigms that permit careful manipulation of input and detailed assessment of what participants are able to learn. Studies using linguistic materials will be contrasted with studies using nonlinguistic materials to further explore the locus and domain-specificity of the learning mechanisms under consideration. The answers to these questions will inform an emerging theoretical framework, constrained statistical learning, intended to elucidate the study of language acquisition and other pressing issues in human learning and development.