論郡社布農語的疑問句構
No Thumbnail Available
Date
2010
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
本文主要是探討台東郡社布農語的疑問句構。筆者除了從田調蒐集之語料中描述介紹郡社布農語的疑問句型外，並以疑問詞疑問句(wh-question)的結構和疑問助詞(Interrogative particles)的句法呈現(syntactic representation)為研究重點，佐以審視疑問詞的分布和探討主語優勢制約(subject-sensitive constraint)現象等。
我們認為郡社布農語的疑問詞問句是一種準分裂句(pseudo-cleft)結構，由句首基底衍生謂語(base-generated predicate)加上無中心詞關係子句(headless relative clause)形成。我們首先對郡社布農語疑問詞問句結構提出兩種分析假設：疑問詞前移分析(wh-fronting analysis)和分裂句構分析(cleft analysis)。從理論及語料實證顯示，分裂句構分析才能適當地解釋郡社布農語疑問詞問句結構。主要論證有以下幾點：第一、疑問詞問句在表面句法結構上相似於準分裂句；第二、疑問詞問句結構和準分裂句結構一樣都沒有遵守承接條件(Subjacency Condition)、無法顯現孤島效應(island effect)，這顯示郡社布農語句首疑問詞並非經由移位衍生而來。而透過運用概化控制(generalized control)，郡社布農語疑問詞的依存關係(wh-dependency)仍可以被建構出來；第三、以未來式標記(irrealis marker) na的句法位置而言，郡社布農語句首疑問詞在分裂句構分析下具有謂語的特質，na可成為其前綴(prefix)，但在疑問詞前移分析下，na卻永遠無法為其前綴，不合語法。此外，我們發現疑問詞mavia為首的問句結構不同於其他疑問詞問句。運用Rizzi (1997)的大句子分裂假說(Split-CP Hypothesis)，我們認為mavia跟其他疑問詞比較起來，在句法上基底衍生於較高的位置，也就是疑問結構(INT)的標示語(specifier)位置。此外，對於所有郡社布農語疑問詞皆須強制(obligatorily)位於疑問句首的現象，我們認為這確切地違背了Keenan (1976)的主語優勢制約理論。
關於郡社布農語的疑問助詞的句法呈現，我們認為這些疑問助詞是基底衍生於CP的中心詞(head)，且帶有強烈特徵的疑問([+Q])和擴展投射原則([+EPP])。而先前研究認為郡社布農語是非問句中的adu是不帶有強烈特徵[-EPP]的疑問助詞；但我們發現郡社布農語是非問句中還有另一個疑問助詞ha可以和adu同時出現於句中，而我們認為ha仍是帶有強烈特徵[+EPP]的疑問助詞，而adu在句法表現上比較像是一個非實現情態副詞(irrealis adverb)。參照Rizzi (2002)整合副詞位置在句子左緣(left periphery)結構的製圖理論(cartographic approach)，以及Cinque (1999)所提出的副詞位階(adverbial hierarchy)，我們認為adu是基底衍生於句子左緣區域(left-peripheral domain)非實現情態副詞修飾語結構(MODirrealis)中的標示語位置。
The thesis investigates the interrogative constructions in Isbukun Bunun spoken in Taitung Prefecture. Specifically, we focus on the construction of wh-questions and the syntactic representation of interrogative particles. We claim that Isbukun Bunun wh-questions are actually a manifestation of a pseudo-cleft structure, where the sentence-initial wh-word is a base-generated predicate, followed by a headless relative clause. Two possible analyses of the construction of Isbukun Bunun wh-questions are provided: the wh-movement/fronting analysis and cleft analysis. We argue for the cleft analysis and against the fronting analysis for the following reasons: First, wh-questions structurally parallel pseudo-clefts on the surface. Second, violation of Subjacency Condition in wh-questions proves that the sentence-initial wh-words are not derived through wh-movement as those in English. The lack of island effect can be accounted for with Generalized Control Rule. Third, in terms of the syntactic position of the irrealis marker na, the sentence-initial wh-words are characteristic of a predicate. Among the Isbukun Bunun wh-words, it is found that mavia-initiated wh-questions manifest distinct syntactic structure. Adopting Rizzi’s (1997) Split CP Hypothesis, we propose that mavia is syntactically base-generated in a higher position, namely, the [Spec, INT], compared to other wh-words. The obligatorily sentence-initial Isbukun Bunun wh-words are a ‘real exception’ to Keenan’s (1976) subject-sensitive constraint. We also propose that Isbukun Bunun interrogative particles are all base-generated at C0, bearing the strong feature [+Q] and [+EPP]. Besides, the previously studied interrogative particle adu, while co-occurring with another interrogative particle ha in yes-no questions, behaves more like a irrealis adverb. Making reference to Rizzi’s (2002) cartographic approach to the left periphery integrated with the position targeted by adverbial phrases, and Cinque’s (1999) adverbial hierarchy, we argue that adu is base-generated in the Spec of MODirrealis head in the left-peripheral domain.
The thesis investigates the interrogative constructions in Isbukun Bunun spoken in Taitung Prefecture. Specifically, we focus on the construction of wh-questions and the syntactic representation of interrogative particles. We claim that Isbukun Bunun wh-questions are actually a manifestation of a pseudo-cleft structure, where the sentence-initial wh-word is a base-generated predicate, followed by a headless relative clause. Two possible analyses of the construction of Isbukun Bunun wh-questions are provided: the wh-movement/fronting analysis and cleft analysis. We argue for the cleft analysis and against the fronting analysis for the following reasons: First, wh-questions structurally parallel pseudo-clefts on the surface. Second, violation of Subjacency Condition in wh-questions proves that the sentence-initial wh-words are not derived through wh-movement as those in English. The lack of island effect can be accounted for with Generalized Control Rule. Third, in terms of the syntactic position of the irrealis marker na, the sentence-initial wh-words are characteristic of a predicate. Among the Isbukun Bunun wh-words, it is found that mavia-initiated wh-questions manifest distinct syntactic structure. Adopting Rizzi’s (1997) Split CP Hypothesis, we propose that mavia is syntactically base-generated in a higher position, namely, the [Spec, INT], compared to other wh-words. The obligatorily sentence-initial Isbukun Bunun wh-words are a ‘real exception’ to Keenan’s (1976) subject-sensitive constraint. We also propose that Isbukun Bunun interrogative particles are all base-generated at C0, bearing the strong feature [+Q] and [+EPP]. Besides, the previously studied interrogative particle adu, while co-occurring with another interrogative particle ha in yes-no questions, behaves more like a irrealis adverb. Making reference to Rizzi’s (2002) cartographic approach to the left periphery integrated with the position targeted by adverbial phrases, and Cinque’s (1999) adverbial hierarchy, we argue that adu is base-generated in the Spec of MODirrealis head in the left-peripheral domain.
Description
Keywords
疑問句構, 郡社布農語, 疑問詞問句, 疑問助詞, 準分裂句, 疑問詞移位, 基底衍生, 承接條件, 大句子分裂假設, 主語優勢制約, 製圖理論, 副詞位階, interrogative construction, Isbukun Bunun, wh-question, interrogative particles, pseudo-cleft, wh-movement, base-generated, Subjacency Condition, split CP hypothesis, subject-sensitive constraint, cartographic approach, adverbial hierarchy