基於語料庫的漢語違實句研究

A Corpus-Based Study of Counterfactuals in Mandarin

本文的研究是為了解釋漢語違實句的工作機制。通過對所得語料進行統計分析可以得出以下結論：（一）除了語用因素，漢語違實句與諸多詞匯語法特徵關聯，如否定、情態副詞、祈願語氣、過去時制、第一人稱代詞、指示代詞等，語用干擾度受這些特徵出現數量的制約；(二)不同假設連詞具有不同的違實生成能力；（三）漢語與英語在違實句的使用頻率和標記策略上有明顯差異。本文的語料來源於六個語料庫，分別是：蘭卡斯特漢語語料庫、UCLA漢語書面語語料庫、最新漢語文本語料庫、布朗家族語料庫（英式英語）、布朗家族語料庫（美式英語）和英漢雙語平行語料庫。

This research aims to discover the working mechanisms for counterfactuals in Mandarin. Following statistical analyses, several key findings emerged. First, besides pragmatic influences, Mandarin counterfactuals are found to be positively correlated with negation, emphatic modals, optative mood, past-oriented temporality, first person pronoun, demonstratives, and so on. The strength of pragmatic influences is subject to the variation of features present in the sentences. Second, different hypothetical conjunctions work quite differently in their ability to generate counterfactuality. Third, some significant differences have been found both in the frequencies of counterfactuals and in marking strategies between Mandarin and English. My empirical findings were confirmed by a search through six corpora, namely the Lancaster Corpus of Mandarin Chinese, The University of California at Los Angeles (UCLA) Corpus of Written Chinese, Texts of Recent Chinese (TORCH), CLOB corpus (Brown family, British English), CROWN corpus (Brown family, American English) and The English-Chinese Bilingual Parallel Corpus.