第二语言句法自动化的指标测评研究  

Assessing Measures of Second Language Automatization in Sentence Processing

作　　者：杨晓峰[1] 张晓鹏[1] YANG Xiaofeng;ZHANG Xiaopeng(Xi'an Jiaotong University,Xi'an 710049,P.R.China)

机构地区：[1]西安交通大学,陕西省西安市710049

出　　处：《现代外语》2024年第3期357-369,共13页Modern Foreign Languages

基　　金：国家社科基金项目“二语构式动态发展与概率性学习机制研究”(20BYY084)的阶段性成果。

摘　　要：本文比较了二语句子加工速度(反应时)和加工稳定性(变异系数)对自动性语言知识使用的预测能力,藉此推断加工速度和加工稳定性对第二语言句法自动化测量的有效性。中低和中高水平的两组中国英语学习者完成了句子迷宫任务、口头诱导模仿任务和限时语法判断任务。结果显示:1)中低和中高水平被试的句子加工技能呈现为非自动化和自动化两种状态;2)中高水平组的变异系数是限时语法判断测试成绩的显著预测变量;3)两组被试的反应时和变异系数对其口头诱导模仿测试成绩的预测未能达到显著水平。研究结果说明,与加工速度相比,加工稳定性可更有效反映二语句法的自动化水平。This study tested the effectiveness of processing speed and processing stability in measuring automatization in L2 sentence processing by comparing the predictability of processing speed(reaction time,RT)and processing stability(coefficient of variation,CV)over the use of automated L2 knowledge.Chinese English learners of intermediate-low and intermediate-high proficiency levels completed tasks involving sentence maze,timed grammaticality judgment,and elicited imitation.Results revealed that:(1)the L2 sentence processing skills of participants at both proficiency levels were categorized into two states:non-automated and automated;(2)CV rather than RT of the intermediate-high proficiency participants,could significantly predict the use of automated knowledge measured by the timed grammaticality judgment task;(3)RT and CV of both groups could not significantly predict the use of automated knowledge measured by the elicited imitation task.Results suggest that CV is more effective than RT in measuring L2 syntactic automatization.

关 键 词：二语句法自动化 自动性语言知识 变异系数 反应时 

分 类 号：H319[语言文字—英语]