词语混淆中母语影响的综合性探证方法——语料库、语言测试、回顾性访谈的三角检测

The detection and validation of L1 influence on L2 confusable words: Triangulating corpus data,language tests and retrospective interviews

以英语背景的汉语第二语言学习者(learning Chinese as a second language,以下简称CSL)为例,对语料库、语言测试和回顾性访谈数据进行三角检测,探讨词语混淆中母语影响的探查和验证方法。基于中介语语料对英语者词语混淆情况进行量化调查,并对比韩、日、蒙、印尼母语CSL学习者的词语混淆分布,探查出93组英语者特异性混淆词语。语言测试以这些词语为材料,对比英语和其他四种母语CSL学习者的词语使用。结果显示,英语者在其特异性混淆词语上的混淆率显著高于其他母语者,在非特异性混淆词语上的混淆率则较低。回顾性访谈更进一步直接证明英语者的词语混淆受到其母语的影响。研究除了讨论语料库探查法在二语词汇研究中的效度和局限,也呈现了多种数据互证互补的优势,为二语词汇习得中母语影响的研究提供了一种综合性研究方法。

Taking English speaking CSL learners as an example, the study investigated the L1 influence on L2 confusable words through triangulating corpus data, language tests, and retrospective interviews. The computation and comparison of the lexical errors from five learners’ written corpora detected 93 pairs of English-speaker-particular confusable words. CSL learners from five different L1 backgrounds(including English) participated in a mandatory fill-in-the-blank task. The results showed that English speakers made more errors than non-English speakers on the confusable words from the 93 pairs, but made fewer errors on the confusable words that are particular to non-English speakers. The retrospective interview confirmed that the reasons were closely related to the influence of English. The overall findings were discussed concerning the limitation of corpus data and the advantage of data triangulation in the mixed-method research of L2 vocabulary acquisition.