台語全稱詞「逐」與「每」的分佈及認可條件

Abstract

本文旨於探論台語中的兩個全稱量化詞——「逐」（ta̍k）與「每」（muí）。「逐」與「每」的研究十分稀缺，而華語「每」的相關討論則多聚焦於其與算子「都」的互動。本文選採不同的取徑，特別檢視了「逐／每」結構未和「攏」（lóng）搭配的情形。我們主張主語及賓語ta̍k/muí-NP都帶著一項無法解讀的焦點特徵，需要帶有焦點特徵的認可條件（licensor）來完成對協（agreement）。我們以焦點結構理論（cf. Rooth (1985, 1992, 1996); Shu (2019)）來解釋包含「攏」、其他副詞成分以及依賴性不定詞等認可條件的共通性，並主張「攏」為一觸發最大化預設（maximize presupposition）（見Liu (2021)）的焦點敏感成分。我們也比較「逐」與「每」的語意差別，同作為全稱量化詞，「逐」易產生集體性解讀而「每」易產生分配性解讀。本文也對「逐」與「每」在英語及華語中的對應成分進行簡單比較。不同於each而較類近於every，「逐」和「每」無法單獨出現在浮游（floated）或並列（binominal）位置。另一方面，在副詞與定詞用法上，台語的「逐」較趨同於華語「每」而台語的「每」則較趨同於華語「每一」。This thesis aims to investigate ta̍k and muí—two universal quantifiers in Taiwanese. Research on ta̍k and muí is sparse, and related discussion on Mandarin mei mostly focuses on the interaction between mei and the operator dou. This thesis adopts a different approach by also looking into ta̍k/muí-constructions without lóng.I argue that both subject and object ta̍k/muí-NPs are equipped with an uninterpretable focus feature [+ u F], and thus need to AGREE with the overt licensors with the [+ i F] feature to execute the uninterpretable feature deletion. I apply focus construction (cf. Rooth (1985, 1992, 1996); Shu (2019)) to explain the shared nature of the licensors, including lóng, other adverbial elements as well as dependent indefinites. The puzzle of ta̍k/muí-NPs and lóng is solved by arguing that lóng is another focus-sensitive component that triggers maximize presupposition (Liu (2021)). I also compare the semantics of ta̍k and muí, stating that although both are universal quantifiers, ta̍k tends to allow a collective reading while muí prefers a distributive one.Lastly, a brief comparison of ta̍k/muí and their counterparts in English and Mandarin is provided. Unlike each, every does not appear in floated or binominal positions, which is similar to ta̍k and muí. In contrast, mei aligns with ta̍k and mei yi more with muí in that their adverbial and determiner performances correspond with each other more.