臺灣閩南語方位詞組的語義網絡：以ti7、tiam3、tua3及其搭配詞為例

Abstract

本研究從認知語義學的角度探討臺灣閩南語的三個近義的ti7、tiam3、tua3方位詞組的語義網絡與使用分佈。由於三個方位詞組「存在於某處」的語義一般來說在使用中可以交替互換，它們在這個語言中的多重功能與使用分佈是一個值得探討的問題。藉由觀察取自於台語文詞彙檢索語料庫中以使用為本的語料(Iunn & Lau, 2007)，我們發起一個二元的討論以細究這三個方位詞組其多義性的動因以及它們身為近義詞組在臺語中的使用分佈。 為了探究三個方位詞組的多義，本研究採用原則性多義(Principled Polysemy) (Tyler & Evans, 2001, 2003) 來界定它們的核心語義(proto-scene)及其延伸語義。我們進一步以 Radden & Panther (2004) 所提出的語言動因(linguistic motivation)中的動因(motivational factors) 說明三個方位詞組從核心語義至其衍生語義的語義擴張過程。在三個個案當中，我們發現它們的語義延伸主要是透過概念基模(image schema) (Johnson, 1987; Lakoff, 1987; Mandler, 1992)、識解(construal) (Langacker, 1987, 2008)、概念隱喻與轉喻(conceptual metaphor/metonymy) (Lakoff & Johnson, 1980)等動因所促發的一系列概念基模轉化(image schema transformation) (Lakoff, 1987; Dewell, 1994)，是為空間認知過程的產物。根據三個方位詞組的多義語義，我們提出其個別的語義網絡。 為了仔細描述這三個近義的方位詞組在臺語中的使用分佈，我們首先以語義圖譜(Semantic Map) (Haspelmath, 2003) 的方法為出發點來比較三個方位詞組的多義語義網絡在這個語言中的分佈。這個分佈以簡明的方式呈現它們於語義延伸方向與使用上分佈的不同。基於其中的差異，我們進一步以語料中的頻率(frequency)來思考三個近義詞組的使用分佈。本研究從整體的角度來調查其意義(meaning)、構式(construction)及搭配詞(collocation)的使用分佈，再從個體的角度來檢視各個語義與構式、搭配詞之間分布的對應。我們發現，從整體來看，三個近義詞組各自擁有它們在語義、空間維度、構式、搭配詞等方面的使用傾向。從個體來看，各個語義及其構式和搭配詞間對應上的分佈亦提供了對於各個語義與各個方位詞組其以使用為本的使用傾向。 除此之外，本研究也著眼於可能促使tiam3和tua3由動詞語法化為介詞的使用分佈，藉由串連在共時語料中所發現它們的一些語法化的跡象，我們嘗試著推測它們個別的型態句法以及語義轉變的可能過程。其後，我們統整促成三個詞彙多義狀態的動因分佈，並發現認知動因(cognitive motivation)為造成它們語義延伸差異的決定因素。我們基於這些以使用為本的分佈，對於它們的多義在語意圖譜上的分佈，提出一個更全面的描述。考慮到Lien (2003)與Chiu (2014)注意到三個介詞一般來說是可互換使用的，而Lien進一步從語氣(mood)的角度指出它們使用上的差異，本研究則從以使用為本的角度揭示它們在臺語中身為近義方位詞組並不完全可以互換使用。 最後，本研究以ti7方位詞組的語義為例，提出一個三維語義網絡模型(three-dimensional semantic network model)來適切的呈現其被歸類在不同語義群的語義間的概念近似性而以語義/概念連續性的形式分佈。

The present thesis aims to investigate the semantic networks and the usage distributions of the three near-synonymous ti7, tiam3, and tua3 locative phrases in Taiwan Southern Min (TSM) from a cognitive semantic perspective. As all of them generally seem to interchangeably serve as the sense “existing in/at a certain location” in use, it is an intriguing inquiry regarding how they function and distribute in this language. By means of observing the usage-based data of the three locative phrases retrieved from the Taiwanese Concordancer (Iunn & Lau, 2007), we launch into a twofold discussion to scrutinize the motivation of their polysemy into the form of the semantic network and the usage distributions of them as near-synonymous counterparts in TSM. To probe into the polysemy of the three near-synonymous locative phrases, the present thesis adopts the Principled Polysemy approach (Tyler & Evans, 2001, 2003) to distinguish their primary sense (proto-scene) and distinct derived senses. We further apply “motivational factors” proposed by Radden & Panther (2004) to elaborate their semantic extensions from the primary sense to other derived senses. Our findings reveal that, among the three cases, the sense extensions are primarily motivated through image schema (Johnson, 1987; Lakoff, 1987), construal (Langacker, 1987, 2008), and conceptual metaphor and metonymy (Lakoff, 1980), respectively as experiential, perceptual, and cognitive motivations, which result in image schema transformation (Lakoff, 1987; Dewell, 1994) as a consequence of the cognitive operation underlying spatial cognition. According to the motivation of their polysemy, we propose a semantic network for each locative phrase in the form of meaning categorization. To delineate the usage distributions of the three near-synonymous locative phrases in TSM, we at first compare their polysemous semantic networks through the semantic map approach (Haspelmath, 2003) as a starting point to identify their distribution in this language. This distribution exhibits that they differ from one another in terms of their extensional directions and use in a concise way. On the basis of the differences on the map, we further consider the usage distributions of the three near-synonymous counterparts from various aspects by way of frequency in a usage-based view. The present thesis surveys the differences of the distributions in meanings, constructions and collocation in general, and in the correspondences among each sense, constructions and collocation in particular. Our findings show that, in general, the three near-synonymous locative phrases have their own propensities for use with respect to sense, spatial dimensionality in the construal of space, types of construction, and collocation. In particular, the distribution of each sense and its correspondence with types of construction and collocation offers usage-based tendencies for the use of each sense and meanwhile each locative phrase. In addition, we ponder over the potential constructional distributions of tiam3 and tua3 which may yield grammaticalization from a verb to a preposition which is an outcome of genetic motivation induced by communicative and cognitive motivations. By interlinking the indications of their grammaticalization observed in our synchronic data, we attempt to conjecture about the possible process of the morphosyntactic and semantic changes of each preposition. By and large, we encapsulate the distribution of the motivational factors that motivate the polysemous status of the three locative phrases. It is discovered that cognitive motivation is a determinant for the differences of their sense extensions. Based on these usage-based distributions, we furnish the distribution of their polysemy on the semantic map with a more full-scale account. Although Lien (2003) and Chiu (2014) have noted that the three prepositions are interchangeable in general and Lien further indicates their difference in use from the perspective of mood, the present thesis subsequently discloses that the locative phrases headed by the three equivalent prepositions are not entirely interchangeable in TSM from a usage-based perspective. Last but not least, with the senses of the ti7 locative phrase as a case study, the current thesis proposes a three-dimensional semantic network model to properly demonstrate the conceptual proximity among the senses that are classified into the different clusters and therefore in the form of the semantic/conceptual continuum.