期刊
CURRENT DIRECTIONS IN PSYCHOLOGICAL SCIENCE
卷 22, 期 3, 页码 239-243出版社
SAGE PUBLICATIONS INC
DOI: 10.1177/0963721413476035
关键词
statistical learning; language; infancy; implicit learning
The term statistical learning was originally used to describe sensitivity to conditional relations between syllables in the context of word segmentation. Subsequent research has demonstrated that infants are sensitive to many other kinds of statistical information. The range of statistical learning phenomena presents a challenge to prior theories and models, which have primarily focused on a single aspect of learning. From our perspective, sensitivity to conditional information yields discrete representations (such as words). Integration across these representations yields sensitivity to distributional information. To achieve sensitivity to both kinds of statistical information, we propose a framework that combines processes that extract units from the input with processes that compare across these extracted items. We review the literature on statistical learning to show how these processes map onto prior research, and we discuss how the interaction between these processes gives rise to more complex patterns of learning than either process achieves in isolation.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据