Statistical Learning Across Modalities, Domains and Languages

Abstract

This dissertation establishes a comprehensive line of research that systematically exploresstatistical learning (SL) across modalities, domains, and languages. The first study delves into SL development, exploring SL changes across different ages, linguistic and non-linguistic domains, visual and auditory modalities, and languages (specifically, Chinese and English). In the second study, the focus is shifted to the probabilistic regularities embedded in a specific written language. This study investigated the ability to use word endings as a probabilistic cue to lexical stress among those who acquired a second language (L2) in formal learning settings through years of practice. The third study broadens the scope further, delving into whether beginning English learners benefit from a specialized training program emphasizing the association between word endings and lexical stress. The collective body of research in this dissertation makes significant contributions to the fields of cognitive science, linguistics, and education. Ultimately, the insights gleaned have the capacity to positively impact learners of various ages and diverse backgrounds, and to offer valuable implications for educational practices.