背景: 生醫文獻語意角色標註（Semantic Role Labeling, SRL）是一種自然語言處理的技術，其可用來將描述生物過程的語句以predicate-argument structures ( PASs ) 表示。SRL 經常受限於arguments的unbalance problem而且需要花費許多的時間和記憶體空間在學習 arguments 之間的相依性。 方法: 我們提出一Markov Logic Network ( MLN ) -based SRL之系統，且此系統使用自動化生成之SRL 模板同時辨識constituents與候選之語意角色。 結果及結論: 我們的方法在BioProp語料上來評估。實驗結果顯示我們的方法勝過目前最先進的系統。此外，使用SRL模板後，在時間及記憶體之花費上亦大幅的減少，而且我們自動化生成之模板亦能幫助建立這些模板。我們認為本論文提出之方法可以透過增加新的SRL模板例如：由生物學家手動寫的模板，而得到進一步的提升，而且本方法也為於需要處理大量SRL 語料時，提供一種可能的解法。 / Background: Biomedical semantic role labeling ( SRL ) is a natural language processing technique that expresses the sentences that describe biological processes as predicate-argument structures ( PASs ) . SRL usually suffers from the unbalanced problem of arguments and consuming time and memory on learning the dependencies between the arguments. Method: We constructed a Markov Logic Network ( MLN ) -based SRL system, and the system uses SRL patterns, which utilizes automatically generated approaches, to simultaneously recognize the constituents and candidates of semantic roles. Results and conclusions: Our method is evaluated on the BioProp corpus. The experimental result shows that our method outperforms the state-of-the-art system. Furthermore, after applying SRL patterns, the costs of the time and memory are greatly reduced, and our automatically generated patterns are helpful in the development of these patterns. We consider that our method can be further improved by adding new SRL patterns such as biological experts manually written patterns and it also provide a possible solution to process large SRL corpus.

過往對於情緒分析的研究顯少聚焦在讀者情緒，往往著眼於筆者情緒之研究。讀者情緒是指讀者閱讀文章後產生之情緒感受。然而相同一篇文章可能會引起讀者多種情緒反應，甚至產生與筆者迥異之情緒感受，也突顯其讀者情緒分析存在更複雜的問題。本研究之目的在於辨識讀者閱讀文章後之切確情緒，而文件分類的方法能有效地應用於讀者情緒偵測的研究，除了能辨識出正確的讀者情緒之外，並且能保留讀者情緒文件之相關內容。然而，目前的資訊檢索系統仍缺乏對隱含情緒之文件有效的辨識能力，特別是對於讀者情緒的辨識。除此之外，基於機器學習的方法難以讓人類理解，也很難查明辨識失敗的原因，進而無法了解何種文章引發讀者切確的情緒感受。有鑑於此，本研究提出一套基於語意框架(frame-based approach, FBA)之讀者情緒偵測研究的方法，FBA能模擬人類閱讀文章的方式外，並且可以有效地建構讀者情緒之基礎知識，以形成讀者情緒的知識庫。FBA具備高自動化抽取語意概念的基礎知識，除了利用語法結構的特徵，我們進一步考量周邊語境和語義關聯，將相似的知識整合成具有鑑別力之語意框架，並且透過序列比對(sequence alignment)的方式進行讀者情緒文件之匹配。經實驗結果顯示證明，本研究方法能有效地運用於讀者情緒偵測之相關研究。 / Previous studies on emotion classification mainly focus on the writer's emotional state. By contrast, this research emphasizes emotion detection from the readers' perspective. The classification of documents into reader-emotion categories can be applied in several ways, and one of the applications is to retain only the documents that cause desired emotions for enabling users to retrieve documents that contain relevant contents and at the same time instill proper emotions. However, current IR systems lack of ability to discern emotion within texts, reader-emotion has yet to achieve comparable performance. Moreover, the pervious machine learning-based approaches are generally not human understandable, thereby, it is difficult to pinpoint the reason for recognition failures and understand what emotions do articles trigger in their readers. We propose a flexible semantic frame-based approach (FBA) for reader's emotion detection that simulates such process in human perception. FBA is a highly automated process that incorporates various knowledge sources to learn semantic frames that characterize an emotion and is comprehensible for humans from raw text. Generated frames are adopted to predict readers' emotion through an alignment-based matching algorithm that allows a semantic frame to be partially matched through a statistical scoring scheme. Experiment results demonstrate that our approach can effectively detect readers' emotion by exploiting the syntactic structures and semantic associations in the context as well as outperforms currently well-known statistical text classification methods and the stat-of-the-art reader-emotion detection method.

一詞多義是所有語言會出現的共有現象，然而，過往較少探討多義詞整體結構與其各義項間的關聯，也鮮少討論句法結構延伸的原因。本篇研究將探討漢語日常使用頻率較高的移動動詞「跑」的多義性，以中研院語料庫之語料為本，客觀的回歸到語言事實作分析，並參照Talmy（1975, 1985, 2000）提出的移動事件框架，檢視「跑」所體現的概念結構及句法結構間的關聯，試圖解釋造成不及物用法至非常規賓語用法背後所展現的機制，此外，本文根據教育部重編國語辭典、中文詞彙網絡知識庫內「跑」的定義及參照Evans（2005）的原則性多義理論，確立「跑」的獨立義項並探討概念結構中不同元素的變化和「跑」多義性之間的關聯，進一步地說明各義項間的相關性與區別性，最後，參照Lakoff（1987）的放射形範疇結構整理出屬於「跑」的語意網絡圖。 本篇研究結果顯示，「跑」共有四種不同的句法結構，分別為最典型的不及物﹝A.名詞短語+跑﹞句式，由趨向介詞帶出名詞短語的﹝B.名詞短語+處所/趨向介詞+名詞短語+跑﹞句式及﹝C.名詞短語+跑+處所/趨向介詞+名詞短語﹞句式、從不及物用法延伸至非常規賓語用法的﹝D.名詞短語+跑+名詞短語（斜格）﹞句式，各句式映射到的概念結構皆不同，是概念結構內不同要素的變化影響句法結構的改變，人類的自然認知過程使路徑延伸出不同種類，造成「跑」的非常規賓語用法以及延伸出不同於字面義的其他義項。此外，本文確立「跑」的七個獨立義項為「以兩腿交互快速向前移動」義、「往特定目標移動」義、「為某事忙碌奔走」義、「迅速離開、逃走、躲避」義、「兩個以上的參與者競速」義、「交通工具或物體的移動」義及「離開原有的位置」義，各義項的出現與整體句法結構、概念結構及語境有很大的關聯，總結來說，「跑」所體現的概念結構、句法結構及語意三者之間的相互影響是造成多義現象的重要因素。 / This study investigates the prevalently used polysemous motion verb PAO “run” in Mandarin Chinese. The data are extracted from Academic Sinica Balanced Corpus of Modern Chinese. Under the framework of Motion Event Frame （Talmy 1975, 1985, 2000）, the study explicates the mechanism of PAO’s syntactic complexities in a more plausible manner. Four different structures of PAO are detected: the most typical intransitive structure [A. NP+PAO], structures with prepostions [B. NP+P+PAO+NP] and [C. NP+PAO+P+NP], and the unconventional transtive structure [D. NP+PAO+NP（oblique case）]. Each structure manifests different conceptual structures. It is claimed that the extention of Path has caused the usage of unconventional transtive structure and the polysemous phenomenon of PAO. In addition, this paper investigates the relationship between the different senses of PAO and elements in its conceptual structure. The relevance and distinction between the different senses of PAO are built as a radial categorization. PAO is suggested to have seven different senses: move quickly with legs, move to specific places, move for specific reasons, escape, two or more participants race, vehicles or objects move and leave the original location. Each sense is related to the overall syntactic structure, the conceptual structure and different contexts. In conclusion, the interaction between the conceputual structure, the syntactic structure and semantic meanings is a major factor that makes a polysemy.

在英文學術論文寫作上，引述動詞在文獻的重述、引用等方面扮演著重要的角色。透過引述動詞的使用，寫作者能更清晰、有力的陳述其論點。目前雖有許多關於引述動詞在句子中表達評論語氣與修辭功能的探討，但卻少有研究提及短語詞組的使用情形。引述動詞短語詞組的進一步探究能使學生在建構學術論述寫作的同時，對於經常與之共現的搭配字詞、句構有更敏銳的覺察能力。本研究從語意及句式兩方面著手，觀察常與PROPOSE和SUGGEST兩個高頻近義引述動詞共現的字詞與詞組。此二動詞所具有的共通點為：(一)字義上皆可表示「提出某事以作進一步考量」、(二)經常用於降低或弱化陳述句中的肯定語氣。本研究的目的在於釐清這兩個近義引述動詞的特點，並更完善地歸納出它們各自所偏好的短語詞組。 本研究的研究工具為擁有近四十五億詞彙的當代美式英文語料庫(Corpus of Contemporary American English)，蒐集了其中PROPOSE 和SUGGEST於學術寫作中的相關語料，以觀察此二動詞的不同詞形(V-base型、V-s型、V-ed型與V-ing型)在句子中的表現。至於本研究的分析則分成了三個部分：(一) PROPOSE 和SUGGEST的語法模組、(二) PROPOSE 和SUGGEST的搭配中，有生命性或無生命性的主詞在主要語法模組中的分布情形、(三) PROPOSE 和SUGGEST後的受詞語義分類。量化分析之餘，本研究也引進了語料庫中的索引行(concordance line)以進一步展示PROPOSE及SUGGEST出現的語言環境，分析模組在不同語境中的特點。 研究結果顯示此二動詞的用法有許多相異之處。SUGGEST具有強烈明確傾向與that子句共同出現，而PROPOSE和that子句共現的機率與其和名詞片語共現的機率相近；除此之外，PROPOSE較常與有生命性的主詞共現，SUGGEST則是偏好與無生命性的主詞共現。至於受詞方面，我們發現PROPOSE後的受詞，多屬「認知觀點」、「方法」、「規則」等語義類別，而 SUGGEST後的受詞則多與「可能性」、「狀態」、「性質特徵」有關，以上的對比結果使我們發現，在學術英文寫作中，PROPOSE通常表「提供計畫或行動供他人參考決定」，而SUGGEST則通常表「傳達可能的推論或概念」。我們也發現此二動詞若是以不同詞形呈現時，偏好的字詞也不同。舉例來說，[suggest that-clause]偏好與表達「研究結果」意義相關的主詞共現，[suggests that-clause]則多與表「研究文獻」有關之主詞一同出現。整體而言，若我們從功能的角度出發，PROPOSE通常用於表示某人提出、建議某事項，SUGGEST則多用於詮釋解讀本研究結果或先前文獻中所提出之觀點。 本研究結果揭示了近義詞間短語詞組使用上的不同，並且也證實了當動詞以不同詞形呈現時會選擇不同的搭配，產生不同的語境。本研究透過系統化分析近義引述動詞的搭配情形，期許能為語言教學教材設計及未來與動詞共現詞組相關之研究帶來啟發與助益。 / Reporting verbs are important in academic research papers for paraphrasing and reviewing previous studies to support a writer’s positions. While a large number of studies have been carried out to investigate the evaluative potential and rhetorical functions of reporting verbs in citations, comparatively little research has focused on the phraseological patterns of particular common reporting verbs, the exploration of which can be beneficial in raising student awareness of the recurrent associations of words and structures of reporting in academic written discourse. This study aims at examining the syntactic and semantic environments of two frequent near-synonymous reporting verbs, PROPOSE and SUGGEST. According to Hyland (1998a) and Hinkel (2016), PROPOSE and SUGGEST both can mean ‘putting forward something for consideration’ and are frequently applied to mitigate the certainty of a statement (e.g., Hyland, 1998a; Hinkel, 2016). We expect to distinguish the two verbs from each other and offer a more comprehensive phraseological profile of them in academic writing. We used the subcorpus of academic writing in the 450 million-word Corpus of Contemporary American English (COCA) as the source data to investigate the performance of PROPOSE and SUGGEST in V-base form, V-s form, V-ed form, and V-ing form. Based on the corpus, the analysis focused on three aspects: (1) the grammatical patterns of PROPOSE and SUGGEST; (2) the distribution of animate and inanimate subjects of the dominant patterns of PROPOSE and SUGGEST; (3) the semantic classification of the objects of PROPOSE and SUGGEST. In addition to the quantitative methods, qualitative-based concordance line analysis was also implemented to reveal the characteristics of the broader stretches of discourse where PROPOSE and SUGGEST occur. The results show that the two reporting verbs are distinct from each other. While SUGGEST has a rather strong propensity for co-occurring with that-clause, for PROPOSE, the possibilities of co-appearing with a simple noun object and with a that-clause are similar. In addition, PROPOSE is more likely to have animate subjects, whereas SUGGEST prefers inanimate subjects. As for the simple noun objects of the two verbs, a large number of instances occurring with PROPOSE systematically belong to the semantic groups of COGNITION, METHOD, and STANDARD; SUGGEST shows more preference for POSSIBILITY nouns, STATE nouns, QUALITY nouns and RELATION nouns. The inter-comparison of the results of PROPOSE and SUGGEST suggests that, in academic writing, PROPOSE usually carries the sense of ‘to offer a plan or action for others to consider’, whereas SUGGEST tends to mean ‘to communicate or show an idea, which is likely to be true’. Moreover, the differences across four word forms of PROPOSE and SUGGEST were also identified. It was found that [suggest that-clause] prefers subjects referring to research results, but [suggests that-clause] occurs with the subjects referring to articles or studies on a particular topic more frequently. Overall, in terms of functions, PROPOSE is mainly used for describing the action of offering something as a choice for people to think carefully; SUGGEST, in contrast, serves to interpret a research finding or present an argument suggested by previous studies and literature. The present study sheds light on the phraseological difference between synonyms. It has also proved that different word forms of the same lemma have different choices of collocations and phraseologies. The research findings will contribute to the teaching and research of English for academic purposes since they provide a systematic analysis of the different habitual collocations of two frequent and similar reporting verbs in research articles. We believe this study will bring some insights to the designs of language teaching materials and can serve as the basis for future studies on the co-occurrence patterns and phraseologies of verbs.

本論文以語意學、認知語言學及心理語言學的觀點，探討中文名詞的多義現象。本篇論文討論詞彙多重意義的語言及心理語言表徵，並研究詞義如何在心理詞彙庫中被擷取。 文中藉由認知語言學者Tuggy（1993）所提出的認知語意架構，及詞彙語意學中「同形異義詞（homonymy）」、「一詞多義（polysemy）」及「模糊意義（vagueness）」等觀念之區辨，探索詞義之本質。本文所談之意義（sense）根據Ahrens（1999）之定義有以下三個特性：（1）意義非轉喻（metonymy）或部分/全體（meronymy）之延伸，但可以是隱喻性延伸；（2）意義之間的延伸關係無法由在同類名詞中直接以規律獲得；（3）除非特意，一詞的不同意義不會在同一語境中同時出現。 針對詞彙辨識歷程的語意效應，本論文有三大研究主題--意義數目效應、意義相關性效應及相對意義頻率效應。各效應之預測如下：「意義數目效應」認為意義數目較多的詞彙在詞彙判斷作業（lexical decision tasks）中比較容易辨識；「意義相關性效應」認為當一個詞的意義之間的相關性較高時，該詞的辨識時間亦會較快；「相對意義頻率效應」則認為一個詞意義之間頻率差異程度越小，該詞越有歧義之特質，也因此越容易辨識。 本研究旨在以實證方式探討這些語意向度在詞彙辨識歷程中所扮演的角色。由受試者提供200個中文名詞的意義並決定意義間的相關性；進行電腦詞彙判斷作業，得到辨識每個詞所需的反應時間；再由統計考驗反應時間來驗證各效應。 本研究結果發現只有「意義數目效應」達顯著水準，意義數目較多的詞彙在詞彙判斷作業中比較容易辨識。此一效應支持心理詞彙庫的隨機觸接模型（random access model）及平行觸接模型（parallel access model）並駁斥「序列搜尋模型（serial access model）」。 / This thesis studies the multiple senses of Chinese nominals from semantic, cognitive linguistic, and psycholinguistic viewpoints. It discusses the linguistic and psycholinguistic representations of a word's multiple senses, and the access of these senses in the mental lexicon. The nature of lexical meaning is examined by discussing lexical semantic notions such as homonymy, polysemy, and vagueness, and by introducing Tuggy's (1993) cognitive linguistic representation of a word's multiple lexical meanings. The "senses" of a word are defined according to the lexical semantic theory of Ahrens (1999) as having three properties: (1) A sense is not an instance of metonymic or meronymic extension, but may be an instance of metaphorical extension. (2) The extension links between two senses cannot be inherited by a class of nouns. (3) Senses cannot appear in the same context (unless the complexity is triggered). This thesis specifically looks into three semantic effects on word recognition, including the word's number-of-sense (NOS) effect, the effect of sense relatedness, and the effect of relative sense frequency. The predictions of these effects are as follows: The NOS effect predicts that words with more senses are recognized faster than those with fewer senses. The effect of sense relatedness predicts that words with more closely related senses are easier to recognize than those with distantly related senses. The effect of relative sense frequency predicts that words with equal sense frequencies are more easily recognized than words with unequal sense frequencies. This research aims at empirically verifying these three effects during the recognition of isolated lexical items. Subjects generate the senses of 200 Chinese nominals, and rate the relatedness among these senses. Lexical decision tasks are conducted to obtain the reaction times required to recognize each stimulus item. The semantic effects are verified by comparing the reaction times of different groups of experimental stimuli. The experimental results confirm only the NOS effect, giving support to the random and parallel access models of lexical access, and refuting the serial access model of the mental lexicon.

本論文旨在探討近義詞的相關議題，特別是要檢視其背後的概念結構與表層的語法功能。透過「方便」與「便利」的研究，本文就這對近義詞在語意上的差異做深入並有系統地分析，並期以此解釋他們在語法層次上不同的表現。 　　以框架語意學(Frame Semantics)為理論架構，本研究發現近義詞「方便」與「便利」會引導出「方便框架」(CONVENIENCE frame)與「不方便框架」 (INCONVENIENCE frame)，而這兩個語意框架雖有相對應的框架成分(frame element)，實質上卻是引發出不同的概念型態。近一步說明這對近義詞在語意上的差別，「方便」與「便利」對於他們所涉入的複雜事件(complex event)持不同的觀點：「方便」主要是專注在結果次事件(result-subevent)並採受恩者(BENEFICIARY)的觀點；而「便利」則較專注在原因次事件(cause-subevent)並採動作者/施恩者(AGENT/BENEFACTOR)的觀點。此外，由於「方便」與「便利」具備了正面、值得嚮往的特質，這也解釋了何以「方便框架」比「不方便框架」有更高程度的目的性(intentionality)。 　　為檢視近義詞在概念上的差異是否會反映於在他們的語法表現上，本研究闡明了「方便」與「便利」的語法功能及其使用分布的情況、以及參與角色(participant role)的凸顯類型(profiling pattern)。結果顯示「方便」與「便利」主要有五種語法功能，即名詞化、修飾名詞、修飾動詞、不及物動詞謂語、及物動詞謂語，最常使用的語法功能為名詞化與動詞謂語(包含及物與不及物)。此五種語法功能皆會突顯某些參與角色，但主要都是突顯了「目的」(PURPOSE)與「手段」(MEANS)；而其他參與角色也會在不同語法功能的使用中被突顯，並且這些突顯類型皆可由「方便」與「便利」在概念上的差異來做解釋。 　　總結來說，本論文闡釋了近義詞「方便」與「便利」在概念上不同的偏好會導致他們在語法上有不同的表現；此外，「方便」與「便利」在參與角色的凸顯類型上亦不相同，這說明了，「方便」與「便利」是屬於不同的構式(construction)。最後，本論文也再次確認了詞彙背後的語意概念會決定其語法的表現。 / The purpose of this thesis is to approach the issue of near-synonyms via the examination of their respective underlying conceptual structures and surface syntactic functions. Specifically, the present study aims to furnish a fine-grained and systematic analysis of the semantic differences between the near-synonymous pair fangbian and bianli that shall better explain their differential syntactic behaviors. 　　Based on the theoretical framework of Frame Semantics, this study found that the conceptual structures of fangbian and bianli are associated with the frames of CONVENIENCE and INCONVENIENCE. While pertaining to a corresponding set of frame elements, the two frames actually prompt distinct conceptualizations. Precisely, fangbian and bianli differ in their perspectivization of the complex event involved: fangbian focuses on the result-subevent and takes the BENEFICIARY’s perspective whereas bianli on the cause-subevent and takes the AGENT/BENEFACTOR’s perspective. In addition, the fact that convenience is desirable and thus typically intended also explains the stronger intentionality involved in the CONVENIENCE frame than in the INCONVENIENCE frame. 　　To investigate whether conceptual differences between the near-synonyms would be manifested in their syntactic behaviors, this thesis further elucidated the syntactic functions and their distribution of fangbian and bianli as well as the profiling of the participant roles in each syntactic function. In particular, five main syntactic functions of fangbian and bianli were identified: nominalization, nominal modifier, verbal modifier, intransitive verbal predicate and transitive verbal predicate; each serves to profile distinct participant roles, mostly PURPOSE or MEANS. Moreover, the profiling of other participant roles can be accounted for by the perspectival distinction between the near-synonymous pair. Finally, the distribution of syntactic functions of fangbian and bianli demonstrated that the usage of the near-synonyms as verbal predicate and nominalization is the most dominant categories. 　　To conclude, this thesis has shown that the conceptual preferences of fangbian and bianli in terms of their perspectivization lead to their different syntactic behaviors. Moreover, the near-synonymous pair also differs in their profiling of the participant roles; in other words, they display distinct profiling patterns and therefore pertain to different constructions. Finally, it still holds for the present study that the semantics of a word drives its syntactic behaviors.