霧台魯凱語的狀語研究

Abstract

作為動詞以及句子的修飾成分，狀語 (adverbial) 在跨語言的表現上呈現出豐富多異的語意內涵以及形態面貌，而本研究意旨探討狀語在霧台魯凱語的句法表現。為了能一窺狀語的豐富內涵，本研究從三個面向逐步切入，本文會先透過句法表現來介紹霧台魯凱語的狀語詞性以及狀語分類，接續會討論且分析狀語在霧台魯凱語所參與的句法結構，最後則會以霧台魯凱語的狀語句法來探究及反思現行主要的狀語理論 (Cinque 1999, Ernst 2002)。 與多數的臺灣南島語相似 (Starosta 1988)，霧台魯凱語的狀語大致在詞性上呈現為動詞並當作句子中的主要謂語，而狀語之動詞特質反映於其可搭配時態和體貌等屈折字綴，以及其可被人稱代詞附著，此二標準皆為霧台魯凱語的典型動詞之表現。而透過狀語與疑問詞的排序和其後方被修飾成分的構詞句法體現，霧台魯凱語的狀語動詞 (adverbial verb) 可進一步分為兩大類，一為結構較低的過程狀語和事件狀語 (process-/event-related adverbial)，二為結構較高的命題狀語 (proposition-related adverbial)。本文發現結構較低的過程狀語和事件狀語在霧台魯凱語當中可參與四種不同的句法結構，分別為連動結構 (serial verb construction)、動後結構 (post-verbal construction)、名物化結構 (nominalization construction) 以及附屬結構 (“subordination” construction)。 在這四種結構當中，本文論證連動結構、名物化結構以及附屬結構皆透過「謂語-補語」(predicate-complement) 的句法結構來呈現「狀語-動詞」(adverbial-verb) 之修飾關係，亦即狀語作為主要謂語來嫁接不同大小的動詞補語結構，如在連動結構裡的輕動詞詞組 (vP)、名物化結構裡的不定子句 (non-finite CP) 以及附屬結構裡的限縮有定子句 (reduced finite CP)。而另一方面在動後結構中，本文論證此結構的狀語應分析為附加語 (adjunct)，附加於所修飾的輕動詞詞組上 (vP)。總而言之，根據句法位置的不同，霧台魯凱語的狀語可體現為動詞，亦可為副詞，而在前者的用法當中，狀語動詞可以嫁接三種不同的補語結構。 在語意層次上，不同的補語結構會在修飾關係中產生不一樣的語意詮釋，如連動結構裡的輕動詞詞組帶有高目的性的詮釋、名物化結構裡的不定子句帶有結果導向的解讀，而附屬結構裡的限縮有定子句則帶有習慣性以及泛指的意涵。透過上述的觀察，本研究認為狀語動詞可視為單一詞項，嫁接不同大小的補語結構，而補語結構的類型會影響狀語動詞的詮釋，因此狀語動詞和其補語有緊密的互動性關係，而本文認為此互動性關係和Ernst (2002) 以語意上的賓語概念來理解修飾範疇和副詞詮釋有平行性。

As adverbials are cross-linguistically diverse in regards semantic denotations and morphosyntactic realizations, this thesis aims to investigate how adverbial modifiers are syntactically reflected in Budai Rukai, one of the Austronesian languages spoken in southern Taiwan. In order to shed light on the adverbial syntax, this thesis launches the discussions on adverbial realizations and classifications, then concentrates on four different adverbial syntactic constructions, and last proposes a reflection on major adverbial theories (Cinque 1999, Ernst 2002) from the perspective of Budai Rukai. In line with most Formosan languages in the literature (Starosta 1988 and among others), Budai Rukai mostly realizes adverbial modifiers as verbal main predicates, evidenced by the attachment of TAM markings and personal clitics. Based on the morphosyntactic coding of modified phrases and the ordering with fronting interrogatives, these adverbial verbs can be further classified into two strata, namely, process- and event-related adverbials in a lower structural position and proposition-related ones in a higher position. It is observed that lower adverbials can be involved in four different constructions, including the serial verb construction, the post-verbal construction, the nominalization construction, and the “subordination” construction. Among the four constructions, this thesis argues that the serial verb construction, the nominalization construction, and the “subordination” construction are all involved with the predicate-complement structure in realizing adverbial modification. Namely, adverbial verbs serve as main predicates taking different sizes of verbal complements, such as vP (serial verb construction), non-finite CP (nominalization construction), and reduced finite CP (“subordination” construction). On the other hand, in the post-verbal construction where adverbials follow action verbs, adverbials are demonstrated to serve as adjuncts modifying a vP. In sum, the same sets of adverbials in Budai Rukai, depending on the syntactic positions, can appear as verbs or adverbs, the former of which can moreover take three distinct types of verbal complements. In the semantic level, different syntactic sizes of verbal complements contribute to various semantic nuances of adverbial modification, such as high intention for the serial verb construction, emotive/result reading for the nominalization construction, and habitual/episodic reading for the “subordination” construction. Through syntax-semantics correspondence, it is discovered that adverbial verbs, as single lexical entries, establish an interactive relationship with their verbal complements, which is parallel to Ernst’s (2002) proposal where modifiees are discussed in the “object” sense, which influences the interpretations of modifiers.