Based on the definition and assessment methods proposed in Chapter 2, an analysis of accuracy at the phonetic, spelling, lexical, syntactic, semantic, and textual levels is given, together with an examination of the impact of extra-linguistic factors, such as the time and place of translation, initiator's policy, and translators. The last section defines the texts for the present study. / Chapter 1 is a review of the literature on translation quality assessment in China and in the West. / Chapter 2 reviews the study of accuracy in scientific and technical translation, including an introduction to accuracy, research on accuracy, and definition of accuracy, which involves completeness in information transfer, consistency in terminology, grammaticality in the encoding of the target text, and methods of assessing linguistic accuracy. / Chapter 3 proposes a new theoretical framework and presents texts for the study of accuracy. Apart from complete transfer of information, consistency in terminology, and grammaticality in the encoding of the target text, accuracy should also be assessed by the function of the target text and the impact of extra-linguistic factors. / Chapter 4 is a contrastive analysis of the source and target texts on the basis of the above theoretical framework. This analysis includes features of the source and target texts, methods in translating terms in traditional Chinese medicine, methods in assessing semantic accuracy, types of Chinese medicine formulas and their methods of translation, types of errors in the target texts, and the impact of extra-linguistic factors on translation. / Chapter 5 discusses methods of assessing accuracy with the use of computer-aided tools, illustrating them with examples and offering explanations on the terms and statistics. The software used are Wordsmith 3.0 (Scott, 1998) and MonoConc Pro 2.0 (Barlow, 2000). / Chapter 6 examines computer-aided assessment of the translation of The Pharmacopoeia with a more comprehensive error analysis of the target texts. It is shown that compared with human analysis, computer analysis has a higher speed in search, a broader range of texts, a faster production of wordlists, and a faster and better calculation of errors. This is a more comprehensive assessment of translation accuracy. It concludes with an analysis of the possible impact of the extra-linguistic factors on linguistic encoding. / Chapter 7 proposes methods to improve accuracy through the construction of the bilingual term bank, and offers suggestions to improve the translation by the construction of a translation memory with the computer-aided tools provided by Trados. A comparison of the new translations with the old ones shows marked improvements in the former. / Chapter 8 concludes with the major findings in this thesis. Firstly, computer-aided assessment of accuracy in scientific and technical translation is faster and broader than human assessment. Secondly, methods for assessing accuracy with computer-aided tools are important in filling up gaps in human assessment. Lastly, computer-aided accuracy assessment has a great impact on other relevant areas. / In scientific and technical translation, accuracy is most important. However, the operative aspect of accuracy assessment still eludes us. This is a thorny issue in translation studies and practice. / The purpose of the present study is to use computer tools, such as Wordsmith 3.0 (Scott, 1998) and MonoConc Pro 2.0 (Barlow, 2000), to help to assess accuracy in the English translation of Chinese pharmacological writings, a specific area of scientific and technical translation. The texts used for this study have been selected from the bilingual versions of The Pharmacopoeia of the People's Republic of China (Beijing: Chemical Industry Press, 2000; The Pharmacopoeia hereafter). / The theoretical framework adopted in this thesis is the translation quality assessment model proposed by Reiss (1971/2000) and Vermeer (1996). The emphasis is on the definition and assessment of "accuracy" so as to provide an operative method to assess accuracy in translation, which is supplemented by a computer-aided approach to accuracy. This is done in the following ways: (1) The creation of a bilingual term bank of 1,500 entries and a translation corpus based on the source texts, target texts and 8 Chinese-English dictionaries of medicine. (2) The categorization of the formula names and linguistic items in the source texts and the definition of accuracy and its assessment methods based on research findings in scientific and technical translation. (3) The quantitative and qualitative analysis of translation errors based on the above definition and methods. (4) The comparison of new translations, produced with the aid of computer tools, with the old ones to demonstrate the validity and applicability of the new approach to accuracy assessment. / This thesis is divided into 8 chapters. / 錢多秀. / 呈交日期: 2005年12月. / 論文(哲學博士)--香港中文大學, 2006. / 參考文獻(p. 198-211). / Cheng jiao ri qi: 2005 nian 12 yue. / Adviser: Sin-Wai Chan. / Source: Dissertation Abstracts International, Volume: 67-11, Section: A, page: 4165. / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. [Ann Arbor, MI] : ProQuest Information and Learning, [200-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. Ann Arbor, MI : ProQuest Information and Learning Company, [200-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstracts in Chinese and English. / School code: 1307. / Lun wen (zhe xue bo shi)--Xianggang Zhong wen da xue, 2006. / Can kao wen xian (p. 198-211). / Qian Duoxiu.

In this thesis work, the potential benefits of Latent Dirichlet Allocation (LDA) as a technique for code clone detection has been described. The objective is to propose a language-independent, effective, and scalable approach for identifying similar code fragments in relatively large software systems. The main assumption is that the latent topic structure of software artifacts gives an indication of the presence of code clones. It can be hypothesized that artifacts with similar topic distributions contain duplicated code fragments and to prove this hypothesis, an experimental investigation using multiple datasets from various application domains were conducted. In addition, CloneTM, an LDA-based working prototype for code clone detection was developed. Results showed that, if calibrated properly, topic modeling can deliver a satisfactory performance in capturing different types of code clones, showing particularity good performance in detecting Type III clones. CloneTM also achieved levels of performance comparable to already existing practical tools that adopt different clone detection strategies.

Cette thèse se situe dans le domaine du Traitement Automatique des Langues et vise à optimiser la classification des documents dans les moteurs de recherche. Les travaux se concentrent sur le développement d’un outil de détection automatique des thèmes des documents (ATDS-fr). Utilisant peu de connaissances, la méthode hybride adoptée allie des techniques statistiques de segmentation thématique à des méthodes linguistiques identifiant des marqueurs de cohésion. Parmi eux, les chaînes de référence – séquence d’expressions référentielles se rapportant à la même entité du discours (e.g. Paul…il…cet homme) – ont fait l’objet d’une attention particulière, car elles constituent un indice textuel important dans la détection des thèmes (i.e. ce sont des marqueurs d’introduction, de maintien et de changement thématique). Ainsi, à partir d’une étude des chaînes de référence menée dans un corpus issu de genres textuels variés (analyses politiques, rapports publics, lois européennes,éditoriaux, roman), nous avons développé un module d’identification automatique des chaînes de référence RefGen qui a été évalué suivant les métriques actuelles de la coréférence. / This thesis in the field of Natural Language Processing aims at optimizing documents classification in search engines. This work focuses on the development of a tool that automatically detects documents topics (ATDS-fr). Using poor knowledge, the hybrid method combines statistical techniques for topic segmentation and linguistic methods that identify cohesive markers. Among them, reference chains - sequences of referential expressions referring to the same entity (e.g. Paul ... he ... this man) - have been given special attention as they are important topic markers (i.e. they are markers of topic introduction, maintenance and change). Thus, from a study of reference chains extracted from a corpus composed of various textual genres (newspapers, public reports, European laws, editorials and novel) we developed RefGen, an automatic reference chains identification module, which was evaluated according to current coreference metrics.

INTRODUÇÃO: A alta prevalência da obesidade infantil em diferentes partes do planeta a posiciona como um dos principais focos de atenção da saúde pública, conhecida a associação dos seus agravos às doenças cardiovasculares e à morte prematura. O objetivo do presente trabalho foi avaliar as intervenções escolares que se utilizam das práticas em atividade física e educação nutricional na antropometria e na pressão arterial de crianças e adolescentes. MÉTODOS: Os artigos foram recuperados por buscas sistemáticas em quatorze bases de dados eletrônicas e por busca manual em listas de referências, com atualização até 30 de setembro de 2012. Dois revisores independentes avaliaram os trabalhos e extraíram os dados. Os trabalhos deveriam atender adequadamente aos seguintes critérios de elegibilidade: população dos 6 aos 18 anos de idade; intervenções comunitárias randomizadas no ambiente escolar, fundamentadas nas práticas em atividade física ou educação nutricional, ou nas duas formas combinadas; presença de grupo controle em paralelo, com seguimento concomitante; descrição de pelo menos um dos desfechos: índice de massa corporal, peso corporal e pressão arterial. A meta-análise foi desenvolvida pelo modelo de efeito randômico, com diferença padronizada entre médias pelo método de Hedges. Também foi realizada a meta-análise de regressão para identificação das fontes de heterogeneidade entre os ensaios, envolvendo as variáveis tempo e tipo de intervenção, faixa etária e qualidade. O grau de heterogeneidade entre os estudos foi verificado pelas estatísticas Q de Cochran e I2, e o viés de publicação foi avaliado subjetivamente por meio da distribuição no gráfico funnel plot. RESULTADOS: Dos 5.899 trabalhos inicialmente recuperados, 140 tiveram seus dados extraídos e 60 remanesceram para a composição das sínteses, assim subdivididas: 12 em atividade física, 8 em educação nutricional e 40 com intervenção combinada. A análise das intervenções combinadas apresentou significância estatística a favor do grupo de intervenção no índice de massa corporal, com a magnitude e variabilidade na medida do efeito em diferença padronizada de -0,14 (IC95%: -0,24 a -0,03; p=0,01; n=29471; I2= 94,4%), ao contrário dos resultados obtidos nas duas intervenções em separado, que não foram conclusivos. No desfecho peso corporal, os resultados foram significantes nas intervenções em atividade física (-0,14; IC95%: -0,27 a -0,02; I2= 7,84%) e nas combinadas (-0,65; IC95%: -1,17 a -0,13; I2= 99,3%). As análises da pressão arterial não obtiveram resultados com significância estatística. O conjunto de todas as intervenções com dados em índice de massa corporal, incluindo 55 estudos, mostrou o resultado de -0,02 (IC95%: -0,03 a 0,00; I2= 94,5%), e nenhuma das covariáveis incluídas na meta-regressão, tempo de intervenção, qualidade metodológica do estudo e faixa etária da população mostrou significância para explicar a heterogeneidade observada. A análise de subgrupos mostrou redução do efeito no estrato por tempo de intervenção curto, de até quatro meses, com estimativa de -0,04 (IC95%: -0,06 a -0,03; I2= 96,6%), e na faixa etária dos seis aos dez anos, com estimativa de -0,23 (IC95%: -0,27 a -0,19; I2= 97,9%). CONCLUSÕES: As intervenções escolares que combinaram atividade física e educação nutricional mostraram redução do índice de massa corporal, ao contrário das intervenções que se utilizaram destes elementos em separado. Os efeitos positivos estão associados aos estudos com menor tempo entre as avaliações iniciais e finais e com as faixas populacionais mais jovens. A alta heterogeneidade observada compromete a validade externa dos resultados e sugere cautela quanto à capacidade de generalização para outras populações. / INTRODUCTION: Because of its high prevalence of in different parts of the planet childhood obesity is one of the main public health issues, with obesity worsening known to be associated with cardiovascular diseases and premature death. The aim of this study was to evaluate school interventions that use physical activity and nutrition education practices in anthropometric measurements and blood pressure in children and teenagers. METHODS: The articles were retrieved via a systematic search of fourteen electronic databases and manual search through reference lists updated until September 30, 2012. Two independent reviewers assessed studies and extracted data. The papers should meet the following eligibility criteria: population aged 6 to 18 years old; randomized community interventions targeting the school environment based on physical activity or nutrition education practices, or a combination of the two approaches; placebo- controlled parallel group and concomitant monitoring; description of at least one of the outcomes: body mass index, body weight and blood pressure. Random-effects meta-analysis was used, with Hedges\'g standardized mean differences. Also, meta-analysis was performed to identify sources of heterogeneity between trials, involving the variables such as duration and type of intervention, age group and quality. The degree of heterogeneity between studies was assessed using Cochran\'s Q statistics and I2 tests, and publication bias was subjectively assessed by a funnel plot. RESULTS: Of the 5,899 papers initially retrieved, 140 were data-extracted and 60 were used in synthesis, as follows: 12 in physical activity, 8 in nutrition education and 40 in combined intervention. Analysis of the combined interventions showed that BMI was statistically more significant in the combined intervention group, with magnitude and variability in the measure of the effect size in standardized difference of (-0.14; CI95%: -0.24 to -0.03; p=0.01; n=29471; I2= 94.4%), unlike the results obtained in the two separate interventions, which were not conclusive. Regarding the body weight outcome, the results were significant for physical activity (-0.14; CI95%: -0.27 to -0.02; I2= 7.84%) and in the combined interventions (-0.65; CI95%: -1.17 to -0.13; I2= 99.3%). Blood pressure analyses did not obtain statistically significant results. The set of all interventions with BMI data (55 studies), showed the result of -0.02 (CI95%: -0.03 to 0.00; I2= 94.5%), and none of the covariates included in meta-analysis, intervention duration, methodological quality of the study and population age could explain the heterogeneity observed. Subgroup analysis showed a reduced effect in the short intervention duration stratum (up to four months), with an estimate of -0.04 (CI95%: -0.06 to -0.03; I2= 96.6%), and in the age group of 6-10 years old, with an estimate of -0.23 (CI95%: -0.27 to -0.19; I2= 97.9%). CONCLUSIONS: The school interventions that combined physical activity and nutrition education led to reduced body mass index, unlike the interventions that used these approaches separately. The positive effects are associated with studies with shorter periods of time between the initial and final evaluations and younger populations. The high heterogeneity observed jeopardizes the external validity of the results and suggests caution in generalizing these findings to other populations.

Text in Afrikaans / Selfmoord is 'n kontroversiele aangeleentheid wat beide omstanders en agterblywendes met onsekerheid vervul. 'n Belangrike rede hiervoor is dat dit 'n onnatuurlike afsterwe is wat teen baie mense se norme en waardes indruis. ldiografiese navorsing is onderneem om die belewenisse van agterblywende gesinslede te ondersoek aangesien hulle die eintlike slagoffers van 'n selfmoord is wat met 'n komplekse realiteit gelaat word. Die doel van die navorsing was om riglyne daar te stel vir die hantering en ondersteuning van hierdie mense. Vanuit 'n in-diepte literatuurstudie, het dit geblyk dat die grootste probleme waarmee agterblywendes gekonfronteer word, wentel random problematiese emosies, veroordeling deur omstanders en belewings van persoonlike onvermoe om te midde van intense rousmart selfaktualisering optimaal te volvoer. 'n Vraelys wat aspekte (socs dit uit die literatuurstudie geblyk het) inkorporeer, is opgestel en aan agterblywendes gegee om te beantwoord. Van die belangrikste elemente wat na vore gekom het, is dat intense verhoudingskomplikasies na 'n selfmoord figureer, dat agterblywendes 'n behoefte het om rouverwerking teen hul eie tempo te doen, dat agterblywendes verkies dat omstanders nie voortdurend die gebeure sal ontleed en soek na oorsake in die teenwoordigheid van agterblywendes nie en dat elke agterblywende persoon se roubelewing persoonlik en uniek is. 'n Hoofstuk is gewy aan 'n wye verskeidenheid terapeutiese tegnieke wat geimplementeer kan word tydens hulpverlening aan agterblywendes. Die studie beskryf ook 'n gevallestudie en die verloop van sestien individuele terapeutiese sessies. In die terapeutiese insette is daar hoofsaaklik gebruik gemaak van hipnoterapie as 'n vorm van ontspanningsterapie waartydens daar gelokus word op problematiese emosies en persepsies wat voortgesette kwalitatiewe selfaktualisering belemmer. In die studie word Opvoedkundige Sielkundige riglyne vir hulpverlening aan agterblywendes, wat benut kan word deur hulpverleners (soos sielkundiges, onderwysers en ander ondersteuningsisteme). verskaf. Die studie bevestig dat die agterblywendes as die werklike slagoffers na 'n selfmoord agtergelaat word, maar dat sinvolle intervensie hierdie persone in staat kan stel om ten spyte van die gebeure weer optimaal te selfaktualiseer. / Suicide is a controversial topic. It causes both bystanders and survivors to feel insecure. An important reason for this is that it is an unnatural death which directly opposes the norms and values of the majority of people. ldiografic research has been undertaken to closely examine the experiences of suicide survivors as they are the real victims who are left to cope with a complex reality after a suicide. The objective of this study was to provide guidelines to effectively manage and render assistance to survivors. From an in depth review of available literature it appeared that troubled emotions, being denounced by bystanders and the inability to proceed with continued qualitative self-actualization are some of the most important problems survivors have to cope with. A questionnaire, incorporating significant aspects reported in literature, was compiled and given to survivors to complete. From their responses certain facets seemed to be problematic. Among the more problematic aspects were complicated interpersonal relations, the basic need to grieve at their own pace, the preference that bystanders avoid analysing the suicide and possible contributing factors (especially in the presence of survivors) and the recognition that each survivor's grief is personal and unique. A chapter has been devoted to a wide range of therapeutic techniques that can be implemented to facilitate the support given to survivors. The study also describes a case study as well as the progression of sixteen individual therapy sessions. Therapeutic intervention was mainly based on hypnotherapy as a form of relaxation therapy during which the focus is placed on problematic emotions and perceptions which prevent continued qualitative self-actualization. Educational Psychological guidelines along which assistance can be rendered by psychologists, teachers as well as other support services are given. This study not only confirms that the survivors are the actual victims of a suicide, but also that it is possible to assist these people to achieve optimal self-actualization through meaningful intervention. / Psychology of Education / D. Ed. (Sielkundige Opvoedkunde)

現今資料大量且快速數位化的時代，各領域對資訊探勘分析技術越趨倚重。而在數位人文中領域中從2009年「數位典藏與數位人文國際研討會」開始，此議題逐漸受到重視，主要目的為將數位文物結合資訊分析與圖像化輔助，透過不同層面的詮釋建構出更完整的文物資訊。 本研究建構一個針對各種中文語料分析的工具，藉由latent semantic analysis、pointwise mutual information、Person’s chi-squared test、typed dependencies distance、word2vec、Gibbs sampling for latent Dirichlet allocation等計算語料中關鍵詞彙關聯強度的方法，並結合分群方法找出可能的主題，最後擷取符合分群結果的相關句子予以輔助人文學者分析詮釋。透過提供各種觀察語料的面向，進而提升語料相關研究學者的效率。 我們利用《人民日報》、《新青年》、《聯合報》、《中國時報》作為實驗與測試的中文語料。且將《新青年》藉由此套工具分析後的結果提供給專業人文學者，做為分析詮釋的參考資訊與佐證依據，並在「2015年數位典藏與數位人文國際研討會」中發表論文。目前我們透過各種中文語料評估工具的效能，且在未來將公開此套工具提供給更多學者使用，節省對於語料分析的時間。 / In recent years, a wide variety of text documents have been transformed into digital format. Hence, using data mining techniques to analyze data is becoming more and more popular in many research fields. The digital humanities gradually have taken seriously since "International Conference of Digital Archives and Digital Humanities" began in 2009. The main purpose of the digital heritage combined with information analysis and visualization could improve the effectiveness of cultural information through different levels of interpretation. In this study, we construct a set of tools for Chinese text mining, calculating associated strengths of collocations work through latent semantic analysis, pointwise mutual information, Person’s chi-squared test, typed dependencies distance, word2vec, and Gibbs sampling for latent Dirichlet allocation etc. The tools employ clustering method to identify the possible topics, meanwhile, the tools will extract the relevant statements according to the clustering results. These clustering and relevant statements contribute and improve the efficiency of humanities scholars’ analysis through providing a variety of observations about the corpora. At the experimental stage of this study, we considered the "People's Daily", "New Youth", "United Daily News", and "China Times" as as the corpora for testing. Among the research, humanities scholars analyzed the "New Youth" by the tools and published a paper in the "2015 International Conference of Digital Archives and Digital Humanities". Currently, we assess the effectiveness of the tools through a variety of Chinese corpora. In the future, we will make the tools freely available on the Internet for Chinese text mining. We hope these time-saving tools can assist in humanities scholars’ study of Chinese corpora.

台灣偶像劇自2001年開始發展至今，已經成為台灣電視劇的主流，在面對高度競爭的收視市場，如何能吸引觀眾收視，行銷是重要的。而在傳播行銷過程中，如何能引起媒體報導加以宣傳、如何能有效使用行銷工具、整合行銷偶像劇，「話題」是重要的。 話題行銷就是將偶像劇產品一致性的訊息設定，建構媒體有興趣報導的題目及內容，也就是媒體話題，透過媒體的報導宣傳，形成閱聽者有興趣的話題，再加以討論、流行、傳播，形成有效傳播產品訊息的行銷。 本研究是以「命中注定我愛你」偶像劇為例，由於「命中注定我愛你」偶像劇創下了台灣偶像劇歷史的最高收視紀錄，分段收視高達13.64%；另一方面本劇亦獲得民國97年度廣播電視金鐘獎的「年度電視節目行銷獎」，本劇在播出期間引起平面媒體爭相報導，共計超過二百篇平面報導、網路熱烈討論，話題熱度高到平均天天平面媒體均加以報導，本劇在台灣偶像劇發展史中，的確具有指標性意義，故以「命中注定我愛你」偶像劇為研究個案。 本研究分別從話題行銷的「策略面」、「戰略面」、「執行面」加以分析，並且從「大眾傳媒通路」及「行銷工具通路」兩個面向，即傳媒通路應用的「報紙」、「電視」、「網路」、「雜誌」及行銷工具中的「活動」、「廣告」、「公關」、「DM」中，執行整合行銷傳播的應用研究， 在話題行銷的「策略面」及「戰略面」分析中是以「傳播模式」及「整合行銷傳播」加以研究，在話題行銷的「執行面」研究，是透過本劇的行銷團隊、產製劇組、媒體記者的實際訪談及相關資料加以整理分析，匯整訪談重點及行銷資料，找到「話題行銷」於「命中注定我愛你」偶像劇行銷的研究發現及結論。 / Since year 2001, Taiwanese idol drama has become the main trend of Taiwanese TV drama show. Marketing is very important in how to attract the audience in a highly competitive TV rating market. However, within the process of media marketing, the “talking point” is very significant because it is how we catch the media’s attention, has them advertise for us and how we can use the marketing implements effectively, and unifies the marketing strategies for the idol drama. “Hot topic marketing” is to set up a certain type of message that makes the peripheral products of the idol drama has its consistency. In this way, it constructs the topic and context that the media is interested. Through media’s reports and advertising, the topic becomes an interesting issue among the audience. When the audience talk about it, it become more popular so it is spread out and the marketing of product selling become even more effective. This research uses the idol drama “Fated to love you’ as an example as a case study. The drama show “Fated to love you” created the highest TV rating record among the Taiwanese idol drama records ever, and the sectional rating was high up to 13.64 ％. Besides, the script also won the prize of “Yearly Best Marketing TV program” in Golden Bell Awards in year 2008. During the broadcasting time, it attracted so many print media to rival for priority to report news about the show. There were over 200 written reports writing about this drama show, and it is also been discussed through the interest ardently. The popularity was so high at the time that print media almost have reports about the show daily. Therefore, the idol drama show “Fated to love you” is very indicative in the development of Taiwanese idol drama show. This research will talk about the “maneuver aspect’, “strategic aspect”, and the “executive aspect” of hot topic marketing and using them to analyze the drama show. Also, through the dimension of using “Mass Media approach” such as newspapers, television, internet, and magazines；“Marketing implements approach” such as propaganda activities, advertisement, relations, and flyers, we can execute the applied research study on integrated marketing dissemination. Within the analysis on the aspect of “maneuver” and “strategy” on buzz marketing, I put emphasis on “Communication Model system” and “Integrated Marketing Communication in particular. I found some discoveries and conclusion on hot topic marketing within the drama show “Fated to love you” through analyzing our marketing team, production group, reporters actual interviews and other related data.

Le travail de thèse présenté dans ce document porte sur la conception d'un service de validation et d'enrichissement d'annotations pour un outil industriel de gestion des connaissances basé sur le langage des Topic Maps (TM). Un tel service nécessitant la mise en oeuvre de raisonnements sur les connaissances, il a été nécessaire de doter le langage des TM d'une sémantique formelle. Ceci a été réalisé par l'intermédiaire d'une transformation réversible des TM vers le formalisme logique des graphes conceptuels qui dispose d'une représentation graphique des connaissances (les TM pouvant facilement en être munie d'une). La solution a été mise en oeuvre dans deux applications, l'une conçue pour la veille médiatique et l'autre pour la promotion de ressources touristiques. Schématiquement, des annotations sont extraites automatiquement des documents selon le domaine concerné (actualité/économie ou tourisme) puis ajoutées à la base de connaissances. Elles sont ensuite fournies au service d'enrichissement et de validation qui les complète de nouvelles connaissances et décide de leur validité, puis retourne à la base de connaissance le résultat de l'enrichissement et de la validation.

<p>The aim of the dissertation entitled Topicalization and<br />Left Dislocation in English and Serbian is to describe<br />and offer a generative account of the syntactic and<br />information-structural properties of topicalization and<br />left dislocation in the languages in question, two<br />superficially similar preposing structures which<br />express the same propositions, but are not felicitous in<br />the same context. The analysis is not contrastive in the<br />sense that we are looking for English-Serbian<br />counterparts or vice versa, but the tertium<br />comparationis are the phenomena of topicalization<br />and left dislocation. The fundamental diagnostics of<br />differentiating between the two variants of left<br />dislocation identified in Serbian is laid out, viz.<br />Hanging Topic Left Dislocation and Contrastive Left<br />Dislocation. The dissertation also explores how<br />informational structuring of an utterance determines<br />contextual choices. The lexicon provides the input to<br />the computational system which by means of features<br />builds structure via phases and also gives rise to<br />displacement. The data related to information structure<br />come from the numeration. The results of the research<br />indicate that the notion of a topic should be<br />deconstructed in a combination of the features [+/-<br />a(nchored), +/-c(ontrastive)], similarly to L&oacute;pez<br />(2009), anchored in the sense of Birner &amp; Ward<br />(1998), and contrastive in the pragmatic sense of Titov<br />(2013), whereby [+/-c] is parasitic on [+a].<br />Topicalization in both English and Serbian, and<br />Contrastive Left Dislocation in Serbian mark topics<br />vii<br />[+a, +c], whereas Hanging Topic Left Dislocation<br />HTLD in both English and Serbian marks topic [+a, -<br />c]. It is argued that adding pragmatic features in<br />addition to formal ones relevant to the derivation by<br />the operation which forms the numeration does not<br />offend the Inclusiveness Condition either. The<br />interpretation of an element marked as a topic is the<br />result of its featural content and its syntactic position,<br />reflecting the interaction of syntax, prosody and<br />pragmatics. Pragmatic features are valued but<br />uninterpretable in the numeration rendering the<br />syntactic object containing it active for syntactic<br />operations. Probe (pragmatic) features trigger<br />Agree(ment), but not displacement. It is an edge<br />feature that drives movement or Internal Merge. It is<br />argued that topicalization in both English and Serbian<br />is generated by movement. The edge feature on C<br />(Force) licenses the TopP if it is required for the<br />interpretation and if it is structurally possible, as<br />observed by Jim&eacute;nez-Fern&aacute;ndez &amp; Miyagawa (2014).<br />Unlike in English, the non-phase head T in Serbian<br />can inherit an edge feature from C and license the<br />TopP in the Spec,TP in those structures which are said<br />to be incompatible with topicalization in English. The<br />head licensing the left-dislocated element is realized<br />as comma intonation, as argued by Emonds (2004). It<br />is argued that Hanging Topic Left Dislocation in both<br />English and Serbian is derived by base-generation of<br />the left-dislocated constituent in its surface position<br />(adjoined to a CP), whereas Contrastive Left<br />Dislocation in Serbian by movement (also to a<br />position attached to a CP). What moves is the<br />resumptive pronoun, and then co-reference is<br />established upon adjoining of the left-dislocated<br />element via the operation Match or Match+Agree of<br />Boeckx (2003), which is the only way not to violate<br />the Inclusiveness Condition. Although both<br />Topicalization and Contrastive Left Dislocation mark<br />contrastive topics, they have different discourse<br />distributions, as confirmed by our corpus, thus this<br />adjoining of a left-dislocated element is justified. In<br />the case of Hanging Topic Left Dislocation, coreference<br />between the left-dislocated element and the<br />resumptive pronoun is established via the operation<br />Match of Boeckx (2003). If the resumptive pronoun is<br />a clitic, it moves to the second position in its intonational phrase triggered by the phonological requirement. Hanging Topic Left Dislocation in both English and Serbian marks referential topics and it is</p><p>also a topic-promoting device in Serbian, as argued on<br />relying on our corpus.</p> / <p>Cilj doktorske disertacije pod nazivom Tematizacija i<br />leva dislokacija u engleskom i srpskom jeziku je da<br />opi&scaron;e i ponudi generativni prikaz sintaksiĉkih i<br />informacijsko strukturnih osobina tematizacije i leve<br />dislokacije u pomenutim jezicima, dveju naizgled<br />sliĉnih struktura za pomeranje reĉeniĉnog elementa u<br />prednje polje koje izraţavaju iste propozicije, ali ne<br />odgovaraju istom kontekstu. Analiza nije kontrastivna<br />u smislu da traţimo englesko srpske ekvivalente ili<br />obratno, već je tertium comparationis pojava<br />tematizacije i leve dislokacije. PonuĊeni su i osnovni<br />dijagnostiĉki testovi za razlikovanje dva oblika leve<br />dislokacije identifikovana u srpskom, naime leve<br />dislokacije odvojene teme i kontrastne leve<br />dislokacije. Disertacija takoĊe istraţuje kako<br />informacijsko strukturiranje iskaza odreĊuje<br />kontekstualne izbore. Leksikon obezbeĊuje ulaznu<br />informaciju sistemu sintaksiĉkih operacija koji putem<br />obeleţja gradi strukturu u fazama i takoĊe dovodi do<br />pomeranja. Podaci vezani za informacijsku strukturu<br />su dati u numeraciji. Rezultati istraţivanja ukazuju na<br />potrebu da se pojam teme ra&scaron;ĉlani na kombinaciju<br />obeleţja [+/-a, +/-c], sliĉno L&oacute;pezu (2009), anaforiĉno<br />u smislu Birnera &amp; Warda (1998), i kontrastno u<br />pragmatiĉkom smislu Titove (2013) pri ĉemu je [+/-c]<br />zavisno od [+a]. Tematizacija i u engleskom i u<br />srpskom jeziku i kontrastna leva dislokacija u srpskom<br />obeleţavaju teme kao [+a, +c], dok leva dislokacija<br />odvojene teme i u engleskom i u srpskom jeziku<br />obeleţava teme kao [+a, -c]. U radu se dokazuje da ni&nbsp;dodeljivanje pragmatiĉkih obeleţja pored formalnih obeleţja relevantnih za derivaciju putem operacije koja formira numeraciju ne naru&scaron;ava uslov ukljuĉenosti. Interpretacija elementa obeleţenog kao tema je rezultat njegove kombinacije obeleţja i njegove sintaksiĉke pozicije, &scaron;to odraţava interakciju sintakse, prozodije i pragmatike. Pragmatiĉka obeleţja su vrednovana, ali netumaĉiva u numeraciji ĉineći sintaksiĉki objekat koji ih sadrţi aktivnim za sintaksiĉke operacije. Upravna (pragmatiĉka) obeleţja uzrokuju slaganje, ali ne i pomeranje. Obeleţje ivice je ono &scaron;to pokreće pomeranje ili interno spajanje. Argumentujemo da je tematizacija i u engleskom i u srpskom jeziku generisana pomeranjem. Obeleţje ivice na upravnom elementu C (Force) dozvoljava TopP ako je to neophodno za interpretaciju i ako je strukturno moguće, kao &scaron;to su formulisali Jim&eacute;nez-Fern&aacute;ndez &amp; Miyagawa (2014). Za razliku od engleskog, nefazni upravni element T (upravni element obeleţja vremena) u srpskom jeziku moţe da preuzme obeleţje ivice od C i dozvoli TopP u Spec,TP u onim strukturama za koje se smatra da nisu kompatibilne sa tematizacijom u engleskom jeziku. Upravni element koji dozvoljava levu dislokaciju se realizuje kao intonacijska pauza, kao &scaron;to predlaţe Emonds (2004). Argumetujemo da je leva dislokacija odvojene teme i u engleskom i u srpskom jeziku nastala generisanjem levo dislociranog konstituenta u mestu realizacije (pridruţenom CP projekciji), dok je kontrastna leva dislokacija u srpskom nastala pomeranjem (takoĊe u poziciju pridruţenu CP projekciji). Ono &scaron;to se zapravo pomera je rezumptivna zamenica i onda se po pridruţivanju levo dislociranog elementa uspostavlja koreferentnost putem operacije uskladi ili uskladi+sloţi Boeckxa (2003), &scaron;to predstavlja jedini naĉin da se ne naru&scaron;i uslov ukljuĉenosti. Iako i tematizacija i kontrastna leva dislokacija obeleţavaju kontrastne teme, one imaju razliĉite diskursne distribucije, &scaron;to je potvrdio na&scaron; korpus, stoga je ovo pridruţivanje levo dislociranog elementa opravdano. U sluĉaju leve dislokacije odvojene teme, koreferentnost izmeĊu levo dislociranog elementa i rezumptivne zamenice se uspostavlja putem operacije uskladi (Boeckx 2003). Ako je rezumptivna zamenica klitika, ona se pomera u drugu poziciju u svojoj intonacijskoj frazi, &scaron;to je&nbsp;uzrokovano fonolo&scaron;kim zahtevom. Leva dislokacija odvojene teme i u englesko i u srpskom jeziku obeleţava referencijske teme i takoĊe je sredstvo unapreĊivanja teme u srpskom, kao &scaron;to se argumentuje na osnovu na&scaron;eg korpusa.</p>

This Diploma thesis deals with some aspects of four chosen Czech daily newspapers referring about the outbreak and early development of events connected with one of the latest significant global crisis - the swine flu, which had broken out in 2009. The theoretical part of the thesis deals with some characteristics of crisis and basic concepts, which might be associated with their media coverage. In the practical part, it deals with the way, how the Czech dailies covered first two weeks of spreading of the disease. At the beginning, the thesis introduces a quantitative profile of the analyzed sample and shows both physical and verbal range of it. The second part is focused on the main topics, which were chosen by all daily newspapers to refer about the swine flu. Those, which did not make it to the main thematic line presented before, are also shortly mentioned in the chapter. In the third part, the thesis follows the main motives appearing in the images that accompanied the news with the chosen topic and also discovers how the disease was "objectified". Finally, language analysis discovering some specific subjective language characteristics referring to a potential personal opinion included by individual authors, or the way how the authors dealt with the coverage of the potential future development...