日語晤談中的省略與言談句法結構：親知標記 to 的出現

Ellipsis and Discourse-Syntactic Structures in Japanese Interview Discourse: The Emergence of the Evidential Marker to

本文以頻率統計為基礎，探討日語晤談中帶補謂語的省略與言談句法結構情境。日語語法主詞一般不出現；帶補謂語亦然。在晤談之中，帶補謂語的省略率相對較低 (32.1%)；但是有帶to 的副詞子句時，省略率即升高。本文發現，表因果、讓步的副詞子句，與to 同時出現的情形最多；表條件、時間的副詞子句則否。這是因為前者提供了親知的訊息，而後者沒有。此外，當副詞子句含有一個帶to 和言說或思想動詞的包孕句時，其後帶to 子句的省略頻率更高。帶to 子句還大量地與重複語句一起出現，明顯帶有訊息親知和肯定的作用。

This study addresses the elision of complement-taking predicates (CTPs) in Japanese interview discourse, and explores the discourse-syntactic contexts which facilitate the elision. A frequency-based approach is chosen to scrutinize the ellipsis and related issues concerning CTP. In Japanese, grammatical subjects are often unexpressed, and even CTPs are omitted in unplanned discourse; the clause contains only the complementizer to. In interview discourse, the overall frequency of CTP elision is relatively low (32.1%); however, when adverbial clauses precede to-marked clauses, the elision of CTP increases. Of the four types of adverbial clauses found in the database used in this study, i.e. cause/reason, concessive, conditional and temporal clauses, cause/reason and concessive clauses most frequently co-occur with to-marked clauses: the former provide the source of evidence for the following to-marked clauses, while the latter contain a fact or notion in spite of which the truth of the main clause is asserted. These two types of adverbial clauses account for 87.4% of all to-marked clauses, and 89.9% of the following to-marked clauses elide the CTP in this clause-linkage. Furthermore, when the preceding adverbial clause contains an embedded clause with to and an overt verb of thinking or saying, the following to-marked clause shows an even higher rate of CTP elision. In addition to these discourse-syntactic properties, to-marked clauses overwhelmingly mark the speaker’s (i.e. 1st person’s) reproduction of his/her own utterance or thought in the past (more than 90%), functioning as an evidential and indicating relatively firm sources of information.