台語「敢」問句的句法

Abstract

本論文探討了台語「敢問句」的句法及問句類型。由於「敢」（kám）總是出現在焦點標記「是」（sī）之前，本研究觀察了焦點標記與句法左緣成份的相對位置，避開了「敢」與位置較高的副詞之間不相容的問題，進而指出「敢」的句法位置應在標句詞詞組（CP），而「敢」之所以也出現於句中相對較低位置，乃句中部份成份「主題化」（topicalization）的結果，經由「主題化」，部份句法成份乃得以脫離焦點範域，此一分析基本上採取了鄭良偉（1983）等人的提議，將主題化視為主題突顯語言（topic-prominent languages）在焦點運作上的主要機制。透過這樣的分析，筆者主張台語的問句成份「敢」應位於語氣詞組（MoodP）之下（見Boya Li 2006，並參見Rizzi 1997, 2001）。

關於問句類型，本論文除了指出「敢」的句法位置不同於正反問句成份，且「敢」在表現上與正反疑問詞組亦很不相同，「敢問句」應非A-not-AB問句的另一種型態（參黃正德1988b, 1991）。此外，參考蔡維天（2008）對於疑問詞組的研究，筆者認為長距離從屬（long-distance scope dependency）與孤島效應（island effects）並不足以支持「敢問句」屬於特指問句（wh-questions）的主張。本研究贊成鄭良偉（1997）、湯廷池（1998，1999）、Crosland（1998）與謝妙玲（2001）等人的看法，主張「敢問句」應屬是非問句類型（yes-no questions）。筆者並進一步指出所有的是非問句都採用了一個共同的架構，在此一架構中，主題化及焦點投射有著密切的合作，就此來看，「敢問句」在是非問句中並非特例。

透過對「敢」與其他是非問句成份表現的觀察，本論文指出「敢」在與焦點的關係上十分不同於其他是非問句成份，進而提議「敢」應是生成於焦點詞組的焦點運算元（focus operator），其後才移入語氣詞組。最後，為了解釋「敢」同時具有中性問句、詰問問句的雙屬性，文中引介了台語中兩個早期的詰問問句成份，並提出「敢」原為中性問句成份，只因做為詰問使用的同源詞「kán」逐漸廢棄不用，因此才形成目前兼具兩種用法的現況。

This thesis deals with the issue of Taiwanese kám questions, including their syntax and question type. I study the distribution of kám on the basis of the rigid order of kám preceding the focus marker sī. By this method, I eschew the incompatibility of kám and high adverbs and reveal that, contrary to what is prevalently held in the literature, kám occupies a relatively high syntactic position in CP domain. On the other hand, I also demonstrate that low positions in which kám appears in a word string are consequences of topicalization as a device for some constituents to escape the scope of focus. This analysis basically follows the proposals in which topicalization is taken as the major focus device in topic-prominent languages (refer to R. L. Cheng (1983) among others). As a result, I pinpoint Taiwanese kám under MoodP (à la Boya Li 2006, also refer to Rizzi 1997, 2001).

Regarding the question type, in addition to proposing that kám occupies a distinct syntactic position, I further demonstrate that kám behaves quite differently from A-not-A sequences and that kám questions cannot be the alternative realization of A-not-AB questions (refer to Huang 1988b, 1991). Based on Tsai’s (2008) studies on wh-elements, I show that the long-distance scope dependency and island effects, observed by Huang (1988b, 1991), do not support the claim that kám questions are wh-questions. Following R. L. Cheng (1997), Tang (1998, 1999), Crosland (1998), and Hsieh (2001), I propose that kám questions are actually yes-no questions. Moreover, I point out that kám questions are not peculiar among their peers by indicating that all yes-no questions share a common structure in which topicalization and the focus projection collaborate closely.

Aside from the general categorization of questions, I also show that kám exhibits a different relationship with focus from other yes-no question markers. Therefore, I suggest that kám is base-generated under FocP as a focus operator before it moves to MoodP. To explain the ambiguous usage of kám, I introduce two old rhetorical question markers in Taiwanese and posit that kám was a pure neutral question marker before the obsolescence of its rhetorical sibling, kán.