http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/14 2026-09-09T03:26:21Z http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/1100 標題: 點亮咖啡香：從認知語言學理解風味的跨感官表達; Brighten Up Your Coffee! Crossmodal Expressions of Flavor in Taiwan Mandarin within a Cognitive Linguistic Framework 作者: Yi-Hsuan Chang; 張亦萱 摘要: 本篇研究從認知語言學的角度出發，探討中文母語人士對風味的感受中，如何以跨感官的方式呈現。風味是直接的嗅覺與味覺體驗，然也普遍認為過於主觀、空虛、抽象，難以具體形塑與表達的感官經驗。而在經由文獻回顧後，可察覺中文母語人士在表達風味體驗時，常藉由非味覺與嗅覺的修辭，作為風味的表徵。本論文以綜觀的角度探討 風味描述中以跨感官用詞描繪的隱喻結構與種類，也從語用的視角檢視跨感官隱喻修辭的目的，進而觀察評鑑者（即說話者）與傾聽者共同建構與共鳴的現象。 本研究以質化的方式分析及檢視語料。對於飲食的感官評析，不少飲食（如紅酒、橄欖油、巧克力、咖啡等等）在製造過程中，必須經由繁瑣的官能鑑定流程，依據統一的風味輪作為量表，始能具體評斷食品的風味優劣。其中，尤以咖啡的風味表達尚未有深入的探討與研究。而咖啡統一的制式化咖啡風味輪（由Specialty Coffee Association of America制定），並無法全然與中文對應，並讓中文使用者理解。因此，為了補足現有研究的不足，本論文從錄像搜集共十小時咖啡官能鑑定過程，並人工轉寫錄影中的風味描述與評析逾兩萬字，作為研究之語料基礎。並提出三個跨感官表達結構 (1) 跨感官隱喻（即共感隱喻Synesthetic Metaphor）(2) 跨感官轉喻（即共感轉喻Synesthetic Metonymy）(3) 跨感官明喻（即共感明喻Synesthetic Simile），以進一步解析風味的表達，與不同知覺間跨感官的聯繫，或與情感、記憶、文化間的微妙關係。研究結果發現，評鑑者運用跨感官隱喻將視覺、聽覺與觸覺中的描繪詞轉而形容風味，跨感官轉喻則為概念式隱喻多是上、少是下(More is up, less is down)環環相扣，而跨感官明喻則藉由意象基模(Image schema)與原形理論(Prototype theory)以期在風味描述中達到說服、感性鋪陳、或感同身受等溝通目的。 再者，本文也提出不同於以往研究中跨感官表現的方向性，並提出可能的修改方向。最後，本論文旨在彰顯跨感官結構在知覺感受表達中所扮演的重要角色，並提倡在語言學中建立知覺與言談(Perception and Discourse)為一研究領域。藉由提出風味表達中的共感隱喻、共感轉喻、及共感明喻的結構，本論文探究嗅覺、味覺與其他知覺如何互動、整合。透過多模態隱喻詮釋，解析語言之於知覺與文化的關聯性，希望能跳脫單純的語言結構與詞彙的脈絡，從認知語言學的角度，進一步釐清人類認知與感受的歷程。; This study aims to explore how flavor is conceptualized crossmodally in perceptions from the perspective of cognitive linguistics by analyzing data collected from coffee cupping events in Taiwan. Although the instinctive senses of smell, taste, and flavor are shared by all, the instrumental convergence of language seems impracticable for conveying such subjective, elusive, and abstract sensation. However, it is through language that our experience of flavor can be reconstructed, evaluated, and expressed (Dyer 2011). The gap between the inexpressible nature of the so-called primitive sensations (i.e., of smell and taste) and language is bridged through figurative expressions like metaphors and similes. Notwithstanding, rarely can we identify the Chinese counterparts of English flavor descriptors, nor is the language of savoring experiences in Chinese well studied. Although coffee cupping involves abundant crossmodal expressions, previous studies have scarcely addressed coffee cupping; instead, wine tasting is a more common topic. To bridge this research gap, the current study conducts a corpus-based investigation on cupping data involving 27,043 words from a 10-hour recording. Based on previous flavor researches (e.g., Paradis, 2013), this study places emphasis on the following three aspects to investigate the highly context-dependent synesthetic expressions (i.e., expressions of crossmodal mappings) found during coffee cupping: crossmodal metaphor (i.e., synesthetic metaphor), crossmodal metonymy (i.e., synesthetic metonymy), and crossmodal simile (i.e., synesthetic simile). However, different from the hypotheses of previous researches, we propose a novel directionality of perceptual transfers in crossmodal interactions. In fact, there is no particular rule that crossmodality in linguistic expressions must obey or violate a certain directionality in terms of the perceptual and conceptual mechanisms within sensory expressions. According to our findings, the crossmodal metaphors featuring interactions across touch, sight, taste, and smell are regarded as synesthetic metaphors; the crossmodal metonymy is in fact transferred from the general conceptual metaphor of MORE IS UP, LESS IS DOWN, and is referred as synesthetic metonymy; the crossmodal associating (i.e. synesthetic simile), analyzed through Image schema and Prototype theory, constructs two pathways of both human cognition and emotion in order to perceptually comprehend and emotionally participate flavor experiences. In sum, three complicated crossmodal formations (i.e., synesthetic metaphor, metonymy, and simile) have been established simply due to the human cognitive ability to demonstrate linguistic representation as a unity of senses (see Marks, 1978). Further, by thoroughly analyzing synesthetic forms and functions, the present study aims to deepen an understanding of the emerging role of crossmodality in flavor expressions through the elaboration of the linguistic mechanisms of flavor expressions. By analyzing the flavor expressions that occur in coffee cupping, the study turns over a new leaf in research efforts on the relations among language, cognition, and perception. 2018-01-01T00:00:00Z http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/4932 標題: 體現與生成詞彙之交會：以台灣總統演講中的身體隱喻為例; When Embodiment Meets Generative Lexicon: The Human Body Part Metaphors in Taiwan Presidential Speeches 作者: Ren-feng Duann; 段人鳯 摘要: 本論文旨在結合體現 (embodiment) 和生成詞彙 (generative lexicon) 兩大理論。藉由分析台灣總統語料庫中，由身體部位名稱構成之隱喻�轉喻，本論文凸顯這兩大理論可互補彼此之不足：透過彰顯意義產生的認知層面，體現彌補了生成詞彙未對意義來源說明的缺憾；生成詞彙，特別是物性結構 (qualia structure)，則藉由解釋身體部位做為隱喻�轉喻的選擇限制，來補強體現。我們分析了台灣總統語料庫中的四個身體部位－血、肉、骨、脈－作為隱喻�轉喻的用法，試圖回答下列問題： (1) 體現 (embodiment) 和生成詞彙 (generative lexicon) 如何互補？生成詞彙中的物性角色 (qualia role) 導致身體部位之隱喻�轉喻的使用？抑或是身體部位做為隱喻�轉喻的使用有助於提取該部位之物性角色？ (2) 物性結構如何限制身體部位之隱喻�轉喻的使用？ (3) 物性結構和生成詞彙機制 (generative mechanisms) 如何影響涉及身體部位的隱喻�轉喻之形成和理解？ (4) 國家領導人如何藉由身體部位名詞來表達政治思維？他們對於身體部位的隱喻�轉喻使用如何反映政治脈絡？ 本論文相信，不同的身體部位，以其功能和特徵，在政治演講中會有不同的用法。不同的政治思維也會反映在身體部位名稱的隱喻�轉喻使用上，而變遷的政治環境也會造成不同領導者對於同一身體部位的不同使用。我們也假設，身體部位之能見度 (visibility) 和功能 (telic) 角色為驅使該部位隱喻和轉喻的使用之最重要因素。 　　 本論文對於概念隱喻、生成詞彙和政治話語等研究皆有所貢獻。在方法上，本論文修改了 Pragglejaz Group於2007年所提出之隱喻辨識程序 (metaphor identification procedure)，為中文語料之隱喻辨識提出了較佳的方法。再者，本論文納入生成詞彙理論，俾便研究者更正確地推導出涉及身體部位的隱喻和轉喻，且有助於這類語言現象的理解。在理論上，本論文建立了物性角色和隱喻之間的相互關聯，同時也預測，身體部位的能見度和功能性為促成其隱喻�轉喻使用的最重要因素。在政治話語研究方面，本論文有系統的分析了台灣總統演講中，身體部位的隱喻�轉喻使用，彰顯了政治脈絡如何影響身體部位詞彙的使用。; This dissertation integrates embodiment with generative lexicon. By analyzing the metaphorically/metonymically used human body part terminology in the Taiwan Presidential Corpus, a representative sample of the Taiwanese leadership rhetoric, we reveal how these two theories complement each other on the one hand, and disclose how the changing political context leads to the discriminated uses of the corporeal terms on the other hand. We argue that the two theories can complement each other: Embodiment strengthens generative lexicon by spelling out the cognitive reasons which motivate meaning generation; and generative lexicon, specifically the qualia structure, reinforces embodiment by accounting for the reason underlying the selection of a particular body part for metaphorization. Choosing to analyze how the four body parts—血 xie ‘blood’, 肉 rou ‘flesh’, 骨 gu ‘bone’, 脈 mai ‘meridian’—behave in the Taiwan Presidential Corpus, this dissertation aims to answer the following questions: (1) How do embodiment and generative lexicon interact? Does the qualia role influence the metaphorical/metonymical use of the body part terms? Or does the metaphorical/metonymical use of the body part terms facilitate the retrieval of the qualia role? (2) What is the significance of qualia structure in constraining the selection of body parts for metaphorical/metonymical use? (3) What is the significance of the qualia structure and the generative mechanisms in the formulation and comprehension of the conceptual pairings involving body parts? (4) How are political ideas conceptualized by the country leadership’s use of corporeal terminology? In other words, how can we establish the association between the activation of certain body parts and a certain political context? This dissertation, built on the potentiality to incorporate embodiment and generative lexicon, investigates the body part metaphors/metonyms used in the leadership rhetoric in Taiwan. We hypothesize that different body parts are activated in different ways in political speeches due to their distinctive features and functions, and the visibility and telicity of a body part are the major reasons why the body part is chosen for metaphorical/metonymical use. Moreover, different political agenda are likely to be reflected in the particular uses of corporeal terms, and the change of the socio-political context should lead to the diverging uses of an identical body part referred to in the speeches. This dissertation will contribute to research on conceptual metaphor, generative lexicon, as well as political discourse. Methodologically, this research, modifying the metaphor identification procedure (Pragglejaz Group 2007), provides a better solution for metaphor identification in Chinese data. With the incorporation of generative lexicon, it furthermore facilitates the researcher to more accurately formulate the conceptual mappings involving body part terms, and to better comprehend metaphorically used body parts. Theoretically, taking generative lexicon into consideration, it establishes correlation between qualia roles and the conceptual mappings. Based on the findings, it also predicts that the visibility and telicity of a body part are the most dominant reasons which activate the choice of a body part for metaphorical/metonymical use. In the light of political discourse, it systematically analyzes how the human body parts are interweaved in the country leadership rhetoric, revealing the influence exerted by political context upon the use of corporeal terminology. 2014-01-01T00:00:00Z http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/102738 標題: 馬來西亞華語語用標記之探討:功能取向研究; Exploring Pragmatic Markers in Malaysian Mandarin: A Functionalist Approach 作者: 黃程君; Tan Jun Ng 摘要: 本論文探討馬來西亞華語中語用標記的使用情形，旨在說明其功能分佈以及在話語中的實際運作方式。儘管語用標記已在華語及其他漢語變體中受到廣泛研究，既有研究多半側重於個別語用功能的討論，較少關注多重語用向度在實際使用中如何同時共現。作為一種使用於多語社會環境中的華語變體，馬來西亞華語在語用標記的使用方面仍相對缺乏系統性的探討。基於功能取向，本研究將語用意義視為具梯度性、多功能性，且受具體語境制約的語用現象。本研究以取自馬來西亞華語YouTube 影片的自然發生口語資料為分析基礎，選取 「咯」、「罷了」、「OK」、「咩」、「喂」 與 「哇老」 六個語用標記進行分析。這些標記並非依其整體出現頻率選取，而是因其在不同說話者與語境中反覆出現，且呈現出較具馬來西亞華語特色、在其他華語變體中較不顯著或未見的語用功能。本研究採取探索式分析設計。量化分析檢視上述語用標記在多個分析向度上的分佈情形，包括認知模態、態度立場、面子管理、語義韻律與話語語用取向，以說明其功能分佈特徵；質性分析則進一步考察這些標記在具體語境中的運作方式，涵蓋其單獨使用以及在部分情況下的標記組合使用。研究結果顯示，馬來西亞華語中的語用標記並不存在固定或一對一的形式—功能對應關係。相反地，單一語用標記往往可在同一語句中同時承擔多重語用功能，其中某些功能在特定語境下較為核心，而其他功能則相對邊緣，其相對顯著性會隨語境而有所變動，而非內在於標記本身。此類具梯度性的語用模式反映了說話者在話語中協商認知立場、評價態度與面子關係的過程，並進一步涉及分析語料中所呈現的社會文化定位。整體而言，本研究提供了對馬來西亞華語 YouTube 話語中語用標記使用情形的系統性描述，並顯示結合分佈觀察與細緻質性分析對於理解語用意義具有重要價值。; This thesis examines the use of pragmatic markers in Malaysian Mandarin, with the aim of understanding both their functional distribution and their deployment in discourse. Although pragmatic markers have been widely studied in Mandarin and other Chinese varieties, prior research has often focused on individual functions in isolation, rather than examining how multiple pragmatic dimensions co-occur in use. Malaysian Mandarin, used within a multilingual sociolinguistic environment, remains comparatively underexplored in terms of pragmatic marker usage. Adopting a functionalist approach, this study treats pragmatic meaning as gradient, multifunctional, and sensitive to local contextual contingencies. The analysis is based on naturally occurring spoken data collected from Malaysian Mandarin YouTube videos. Six pragmatic markers—lo, baliao, OK, meh, wey, and walao—are selected for analysis. These markers are not chosen on the basis of overall frequency, but because they recur across speakers and contexts and exhibit functions that are characteristic of Malaysian Mandarin and less prominent or absent in other Mandarin varieties. The study adopts an exploratory analytical design. Quantitative analysis examines how the selected markers pattern across several analytic dimensions, including epistemic modality, attitudinal stance, face management, semantic prosody, and discourse-pragmatic orientation, in order to account for their functional distribution. Qualitative analysis then investigates how these markers function in specific contexts, both individually and, where relevant, in selected marker combinations. The findings suggest that pragmatic markers in Malaysian Mandarin do not exhibit fixed or one-to-one form-function correspondences. Instead, individual markers may simultaneously perform multiple pragmatic functions within a single utterance. In such cases, certain functions emerge as more prototypical while others remain more peripheral, with their relative prominence varying across contexts rather than being inherent to the marker itself. These gradient patterns reflect how speakers negotiate epistemic alignment, evaluation, and face concerns in discourse, and they also contribute to sociocultural positioning within the analyzed dataset. Overall, this study provides a systematic account of pragmatic marker use within Malaysian Mandarin YouTube discourse and demonstrates the value of integrating distributional tendencies with close qualitative analysis in the study of pragmatic meaning. 2026-01-01T00:00:00Z http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/104446 標題: 類型學特徵值的學習表徵：以 WALS 與 Grambank 為例; Learned Representations of Typological Feature-Values: A Study Based on WALS and Grambank 作者: 潘同樂; Tung-Le Pan 摘要: 類型學資料庫在記錄一項語言的特性之前，就已經先做出了一些決定。WALS 稱華語具有複雜聲調系統的前提是有人已經決定將「聲調」視為一個獨立的比較概念，並只允許無聲調、簡單聲調系統、複雜聲調系統三種可能的答案。這些答案構成一個封閉的集合，且彼此互斥。因此，每一個特徵值的意涵不僅需探究其自身，也需理解其餘選項所代表的意義：許多跨語言比較正是透過透過如此安排完成分析。 以這些資料庫為對象的計算方面研究多半會忽略這層機制：一個具有七個值的特徵被拆解成七道各自獨立的是非題的時候，模型內部將沒有辦法將同屬動詞居末的主賓謂與賓主謂語序連結，也無從得知「簡單聲調系統」與「複雜聲調系統」皆具備聲調系統的特質。這樣的作法在計算上確實便利，但也犧牲了許多資訊。本論文將一個特徵的各個值視為同一比較概念之下彼此競爭的選項，在訓練模型時，提高其中一個值的機率，將同時壓低其餘各值的機率。接著，當這些值在彼此之間具有相對位置之後，由它們所構成的空間究竟是什麼樣貌？ 本研究的分析以 WALS 與 Grambank 兩個資料庫為中心：兩者涵蓋的內容有重疊，設計理念卻各不相同：WALS 將大多數特徵以多值的形式看待，其中絕大多數的儲存格並無資料；Grambank 則以二元問題為主，詢問某現象在某語言是否存在，而約四分之三的儲存格具有資訊。 儘管特徵預測長年來在類型學相關的自然語言處理研究中具有不少的吸引力，本論文決定不將重點置於此。當同語族的語言在訓練的資料佔比稀少時，類別式的模型表現落後於二元模型的基線。不過，隨著資料量累積，前者的表現則逐漸超前：此一現象在 WALS 上表現得最為清楚。這現象可以由兩個資料庫的特質解釋：Grambank 的特徵幾乎全為二元，而在同一個問題只有兩種答案的情況下，兩種模型的運作方式相當接近。 這些值的表徵的整體表現是本研究的核心：在 WALS 中，向量空間能夠還原二元模型所無法做到的類型學特徵內部組織，包括：格位數量或動詞屈折綜合程度等序列式尺度；聲調與定冠詞等當中「缺無」與「存在」（其中包括各種次分類）之間的區隔；由多組有無對立交叉而成的特徵；以及由一系列不同策略組合而成之特徵的成分結構。 在 Grambank 方面，本文檢視的是特徵之間的關係：以一項評估蘊含共性的研究（Verkerk et al. 2026）為基準，凡是通過親緣與地理兩項控制門檻的蘊含共性，在幾何上所獲得的支持明顯強於未通過者；二元空間則針對未通過的共性，給予的分數幾乎與通過的相當。 本文欲探究向量空間的表現能否反映類型學家之建構，以及其是否符合既知的語言行為相關性。當比較概念的特徵值被視為互為競爭的選項時，透過表徵學習，它們便得以量化成為彼此之間距離可比較、可測量的位置。; A typological database records a decision before it is ready to record a fact. Before WALS can say that Mandarin has a complex tone system, someone has to have decided that tone is a single comparative concept allowing three possible answers: no tones, a simple system, a complex one. The answers consist of a closed set and the answers are mutually exclusive. Each means what it means not only by itself but also what the others have to offer, and many analyses on cross-linguistic comparison are reliant on this arrangement. Much of computational work on these databases has overlooked this arrangement: a feature with seven values becomes seven independent yes/no questions, so nothing in the model connects SOV to OSV despite both being verb-final, and nothing indicates that a simple tone system and complex tone systems both count as tone systems. It is computationally convenient to do so, but at what cost? This thesis treats the values of a feature as competitors within one comparative concept, so that, when training, raising the probability score of one value lowers the rest, and then asks what the map of those values look like once they hold positions relative to each other. The analysis is run on two databases: WALS and Grambank. The two contain much overlap with distinct design philosophies: WALS stores most of these features as multi-valued categories and has the majority of its cells empty, whereas Grambank asks mostly binary questions and contains information for roughly three quarters of its cells. Despite taking up much of the spotlight for linguistic typology NLP throughout the years, feature prediction is in fact the least interesting section of this thesis. The categorical implementation trails behind the binary baseline when evidence from related languages is scarce, and overtakes it as the amount of data accumulates: this phenomenon is seen most clearly on WALS. This is explained by the nature of the two databases: almost every Grambank feature is binary, and with only two answers to the same question, the two models work similarly. The overall representation of the values constitutes the core of this work: in WALS, the vector space can recover the internal organization of typological features that the binary model discards or even reverses, including: graded scales such as the number of cases or amount of verbal inflectional synthesis; the split between absence and the subtypes of presence in terms of tone and definite articles; the axes of features formed by crossing multiple presence / absence contrasts; and the component structure of features whose values are combinations of a series of alternative strategies. In Grambank, the relationships across feature boundaries are inspected. Implicational universals that survive controls for genealogical descent and geographic proximity show stronger geometric support than those that fail them, judged against an external study (Verkerk et al. 2026) which applied these controls and evaluated them; the binary space scores weak universals nearly as high as stronger ones. I asked whether the behaviour of the vector space can reflect the categories typologists constructed, and whether they follow known correlations of linguistic behaviour. Treated as alternatives rather than solely yes / no labels, feature values of comparative concepts are quantified as positions at comparable and measurable distances from one another through representational learning. 2026-01-01T00:00:00Z