論漢語「跟」字句: 其句法範疇及語法化

Abstract

本篇論文主旨討論以漢語詞彙 「跟」字為中心語所構成(headed)之三種句法結構。分別為動詞組附加語(comitative VP-adjunct)，伴同結構(comitative construction) 以及並列結構(coordinate construction)。

漢語字詞「跟」的確切句法範疇(syntactic category)迄今無法定義。根據前人研究，「跟」根據其句法行為可定義為動詞，介詞，連接詞或其它。儘管有相當成熟的句法表現與具有共識的句法分析，仍無法充分解釋語料的現象，譬如，併列結構中的第一並列詞組(first conjunct)抽取(extraction) 。此抽取違反了Ross(1967)所提出的並列結構限制(Coordinate Structure Constraint )，說明了此限制並非牢不可破。另一方面，Kayne(1994)的 with/and替換(alternation)分析以及Zhang(2006)對於英語的伴同結構分析皆無法解釋漢語中的伴同結構以及並列結構。因此，我主張本文所提出的句法分析可以解決語料上的發現並且補充理論上的不足。此外，主張的句法分析和「跟」字的語法化過程與內容相當一致。

本文有四個目的。首先，透過跨語言的比較(俄羅斯語及捷克語)和一項句法測驗(約束原則A)，我證明漢語「跟」字為中心語所構成的結構可分成兩種。第一種又可分為兩部分加以說明，動詞組附加語(comitative VP-adjunct)，伴同結構(comitative construction)，皆以介詞「跟」為中心語。第二種為並列結構(coordinate construction)則是以連接詞「跟」為中心語。其次，奠基於Kitada (2007a, 2007b)以及Zhang (2006)的研究，我主張介詞「跟」和連接詞「跟」皆被功能投射(Functional Projection)所支配，而在此投射範疇內，「跟」體現為邊緣特徵(Edge Feature)。此主張說明兩種範疇的「跟」皆有其複雜的內部結構(rich internal structure)，結構的功能中心語觸發 (trigger) 一致（Agree）以及移動(Move)等句法操作(syntactic operation)。此主張有兩個貢獻:介詞「跟」的伴同(comitative)語意與連接詞「跟」的分指(distributive)語意可以在句法推導(derivation)過程中得到論證，以及併列結構中的第一並列詞組(first conjunct)抽取(extraction)獲得理論上之解釋。第三，採用形式語法化(van Gelderen 2004, 2006; Roberts and Roussou 1999, 2003)的方法，我們發現並得知本文所提出的三種「跟」字句法結構皆與其語法化過程相符合。我主張此語法化可透過後續合併原則(Late Merge Principle)連同特徵經濟原則(Feature Economy)獲得理論上的支持以及解釋。最後，本文所提出的主張希望可用於其它語言，因而得到更多的跨語言支持。

This thesis focuses on three constructions headed by GEN in Chinese, VP-adjunct, comitative construction and coordinate construction.

The precise syntactic categories of GEN, as a verb, a preposition, a coordinator or others, have been the subject of extensive inquiry from various approaches and perspectives, with sufficient progress and convergence (Tang 1979, Li and Thompson 1981, Gu 2000, Paris 2008 and Zhang 2008 and 2009, Tang 2010, among others). Nevertheless, given the studies, some empirical problems remain unsolved and await better treatment, such as extraction of the first conjunct out of a coordinate construction, which violates the Coordinate Structure Constraint (Ross 1967). Along the same line, two theoretical accounts, and/with alternation (Kayne 1994) and comitative construction in English (Zhang 2006), fail to cover the comitative and coordinate construction in Chinese. Granted this empirical finding and the theoretical inadequacy, I argue for my proposed analyses that can cope with the inadequacy and complement the previous analyses. At last, the proposed analyses are consistent with the path of grammaticalization of GEN.

I have four goals in this thesis. First, having gained a comparative perspective from other languages (Russian and Czech) and employing one syntactic test (Binding Principle A), I demonstrate that there are two types of constructions headed by GEN, comitative and coordinate, with the former being further split into VP-adjunct and comitative coordination which are base-generated at different positions. They are headed by the prepositional GEN and the conjunctive GEN respectively. Second, assuming Kitada (2007a, 2007b) and Zhang (2007), I argue that a prepositional GEN and the conjunctive GEN are dominated by a functional projection (FP) where GEN is realized as an Edge Feature, suggesting that they have the rich internal structure in which the functional head of the phrase triggers syntactic operations, such as Agree and Move. Along the line of this argument, comitativity denoted by a prepositional GEN and distributivity by a conjunctive GEN can be explained in the derivation. Besides, granted this argument, possible extraction of DP1 out of the coordinate structure headed by the conjunctive GEN, not discussed in the studies, can receive an explanatory treatment. Third, employing a formal approach to grammaticalization (van Gelderen 2004, 2006; Roberts and Roussou 1999, 2003), I demonstrate that the grammaticalization of GEN is consistent with three syntactic configurations I proposed respectively. Two principles are operative, the Late Merge Principle and the Feature Economy. Fourth, it is hoped that the proposed analyses can receive substantial support from other languages.