統計學習機制在語言習得中的角色及其對第二語教學的可能啟發

The Role of Statistical Learning in Language Acquisition and Its Implications for Teaching a Second Language

不論學習第一語或第二語，均涉及多種基本認知機制的運作。學用合一理論 (Usage-based Theory) 指出，語言習得涉及兩組先天認知機制：追蹤規律之「規則尋覓」機制與判斷語用意向的「意圖判讀」機制。「規則尋覓」機制中的「統計學習」機制可追蹤、計算語言單位間的組合機率。透過此機制，學習者找出基本語言單位並歸納出語言的抽象規則。過往研究已證實統計學習機制在第一語或第二語之學習歷程中均扮演重要角色。此觀點視語言為多層次的統計分佈集合體，強調經驗與輸入的重要性。本文回顧「統計學習」之相關研究，探討既有之研究成果對現代語言教學的可能啟發，指出未來值得進一步探討的應用研究方向，嘗試為「因材施教」、「適性教學」探尋具有理論基礎的應用契機。

Language learning, be it first or additional, necessarily requires the support of basic cognitive mechanisms. The usage-based theory of language learning, as represented by Tomasello’s work, stresses that learning takes place while using the language and that two innate cognitive mechanisms support the learning. The pattern finding mechanism tracks and discovers the regularities in the linguistic input. The intention reading mechanism determines the intended meaning of the linguistic input during social interaction. It is believed that the pattern-finding mechanism tracks and discovers statistical regularities that form the basis of the so-called grammar in a language. Human learners are equipped with the statistical learning mechanism to determine if a given linguistic input is grammatical. Past research has found that the statistical learning mechanism plays an essential role in both first and additional language learning and that it is available to both child and adult learners. The present article reviews this research and discusses its implications for applied research in second language acquisition and the potential applications for teaching a foreign/second language.