英語完結性相關結構之第二語言習得

Abstract

本研究主要探討英語完結性相關結構之第二語言習得，亦即結果結構、動質結構和與格/雙賓結構。為了避免潛在的實驗效應，本研究採用兩種實驗題型，其一是文法判斷題，其二是開放式翻譯題。研究議題包括結構本身對於第二語言習得的影響、三種結構間之關係、屬性效應、組別效應以及題型效應。本研究的受試者為一百零四位以中文為母語的高中生。依據其英語程度，分為初、中和高三組。

整體實驗結果顯示，此三種結構在統計上呈現極顯著之相關。此相關性並不受到屬性效應、不同英語程度和實驗效應的影響，呈現以下習得次序：與格/雙賓結構最先習得、動質結構其次，而結果結構為後。回歸分析更顯露出此三種結構間之階層性預測關係：成功習得結果結構會顯著地預測成功習得動質結構和與格/雙賓結構，且動質結構亦會顯著地預測與格/雙賓結構之習得。檢測六大屬性時，亦發現在第二語言習得結構效應優於屬性效應。而單一結構的分析亦指示屬性效應大部分對於第二語言習得並無關鍵性的影響。此外，第二語言程度的因素亦受到了證實；受試者的表現隨著其英語能力的提升而進步。在題型效應方面，受試者在文法判斷題的表現比翻譯題來得好，顯示語言理解能力較語言使用能力先習得。最後，錯誤分析也證實了此三種結構一致的關係。本研究指出，不同的第二語言複雜程度、第一語言轉移以及第二語言語料刺激可能會導致此完結性相關結構之階層性習得次序。The present study aims to probe into the second language acquisition of three telicity-related constructions in English, viz. resultative constructions, verb-particle constructions, and dative/double object constructions. To circumvent potential methodological effects, this study drew on a comprehension task, i.e. a grammaticality judgment task, and a production task, i.e. a free translation task, to garner pertinent data. The issues addressed included construction effects, the relationships of the three constructions, property effects, group effects, and task effects. The subjects of this study consisted of 104 Chinese-speaking senior high school students, and they were further divided into three groups according to their English proficiency levels, i.e. the high-level group, the mid-level group, and the low-level group.

The overall results indicated that the three constructions were significantly correlated (p<.01** throughout). The following acquisition sequence was found across different properties, tasks, and L2 proficiency levels: datives/double objects were followed by verb-particles, which in turn preceded resultatives. The regression analysis further showed a hierarchically predictive relationship of the three constructions: the success in the acquisition of resultatives well predicted the successful acquisition of the other two constructions, and verb-particles were a significant predictor ofdatives/double objects as well. When all the test properties were compared, it was found that construction effects consistently overrode property effects on the L2 acquisition. In the within-construction analysis, it was also shown that properties played a marginal role in influencing the L2 acquisition of the constructions in most cases. Moreover, the variable of L2 proficiency was substantiated; the subjects’ performance improved with the increase in their proficiency levels. With respect to methodological effects, the subjects consistently performed better on the comprehension task than the production task, implying that comprehension might precede production in acquisition. Finally, the analysis of untargeted production data also confirmed the observed relationships of the constructions. This study suggests that different degrees of L2 complexity, L1 transfer, and L2 inputs may contribute to the hierarchical acquisition order of the three telicity-related constructions.