This thesis is focused on testing the so-called unique items hypothesis on Czech language data. Supposed Czech unique items were chosen from lexical units, word-formation phenomena, syntactic structures and language use phenomena. Their frequency in a comparable monolingual corpus of contemporary Czech was established and the differences in frequency were statistically tested. This quantitative research was accompanied by a qualitative probe into the English source texts from which sentences containing selected unique items were translated using an aligned parallel corpus of English-Czech translations. The results reveal a general tendency of unique items to be underrepresented in translated language and a variety of source- language phenomena that underlie unique items usage in the target language.

Die Arbeit setzt sich mit verschiedenen Aspekten der Nutzung korpusbasierter Statistiken in quantitativen typologischen Untersuchungen auseinander. Die einzelnen Abschnitte der Arbeit können als Teile einer sprachunabhängigen Prozesskette angesehen werden, die somit umfassende Untersuchungen zu den verschiedenen Sprachen der Welt erlaubt. Es werden dabei die Schritte von der automatisierten Erstellung der grundlegenden Ressourcen über die mathematisch fundierten Methoden bis hin zum fertigen Resultat der verschiedenen typologischen Analysen betrachtet. Hauptaugenmerk der Untersuchungen liegt zunächst auf den Textkorpora, die der Analyse zugrundeliegen, insbesondere auf ihrer Beschaffung und Verarbeitung unter technischen Gesichtspunkten. Es schließen sich Abhandlungen zur Nutzung der Korpora im Gebiet des lexikalischen Sprachvergleich an, wobei eine Quantifizierung sprachlicher Beziehungen mit empirischen Mitteln erreicht wird. Darüber hinaus werden die Korpora als Basis für automatisierte Messungen sprachlicher Parameter verwendet. Zum einen werden derartige messbare Eigenschaften vorgestellt, zum anderen werden sie hinsichtlich ihrer Nutzbarkeit für sprachtypologische Untersuchungen systematisch betrachtet. Abschließend werden Beziehungen dieser Messungen untereinander und zu sprachtypologischen Parametern untersucht. Dabei werden quantitative Verfahren eingesetzt.

Le traitement de musique existante pour en construire de nouvelle est une caractéristique fondamentale de la tradition musicale occidentale. Cette thèse propose et discute mon approche personnelle au sujet : l'emprunt de fragments de musique à partir de grands corpus (contenant des échantillons audio ainsi que des partitions symboliques) afin de créer une palette de grains organisée par descripteurs de bas niveau. Les paramètres sont gérés par des partitions numériques hybrides. Cette thèse présente également la bibliothèque "dada", qui fournit au logiciel Max la possibilité d'organiser, de sélectionner et de générer du contenu musical grâce à un ensemble d'interfaces graphiques manifestant une approche exploratoire à la composition. Ses modules abordent, entre autre, la visualisation de bases de données, la segmentation et l'analyse des partitions, la synthèse concaténative, la génération musicale à travers la modélisation physique ou géométrique, la synthèse "wave-terrain", l'exploration de graphes, les automates cellulaires, l'intelligence distribuée et les jeux vidéo. Pour terminer, cette thèse traite de la question de savoir si la représentation classique de la musique, démêlée dans l'ensemble standard des paramètres traditionnels, est optimale. Deux alternatives possibles aux décompositions orthogonales sont présentées : des représentations de partitions fondées sur les "grains", qui héritent les techniques de la composition basée sur corpus, et des modèles d'apprentissage automatique non supervisés, fournissant représentations de la musique "agnostiques". La thèse détaille aussi ma première expérience d'écriture collaborative au sein du collectif /nu/thing. / The reworking of existing music in order to build new one is a quintessential characteristic of the Western musical tradition. This thesis proposes and discusses my personal approach to the subject: the borrowing of music fragments from large-scale corpora (containing audio samples as well as symbolic scores) in order to build a low-level, descriptor-based palette of grains. Parameters are handled via digital hybrid scores, in order to equip corpus-based composition with the control of notational practices. This thesis also introduces the dada library, providing Max with the ability to organize, select and generate musical content via a set of graphical interfaces manifesting an exploratory approach towards music composition. Its modules address a range of scenarios, including, but not limited to, database visualization, score segmentation and analysis, concatenative synthesis, music generation via physical or geometrical modelling, wave terrain synthesis, graph exploration, cellular automata, swarm intelligence, and videogames. The library is open-source and it fosters a performative approach to computer-aided composition. Finally, this thesis addresses the issue of whether classical representation of music, disentangled in the standard set of traditional parameters, is optimal. Two possible alternatives to orthogonal decompositions are presented: grain-based score representations, inheriting techniques from corpus-based composition, and unsupervised machine learning models, providing entangled, `agnostic' representations of music. The thesis also details my first experience of collaborative writing within the /nu/thing collective.

In Cognitive Linguistics, a conceptual metaphor is a systematic set of correspondences between two domains of experience (Kövecses 2020: 2). In order to have an extensive understanding of metaphors, metaphoricity (Müller and Tag 2010; Dunn 2011; Jensen and Cuffari 2014; Nacey and Jensen 2017) has been emphasized to address one of the properties of metaphors in language usage: gradience (Hanks 2006; Dunn 2011, 2014), which indicates that metaphorical expressions can be measured. Despite many noteworthy contributions, studies of metaphoricity are often accused of subjectivity (Müller 2008; Jensen and Cuffari 2014; Jensen 2017), this is why this study uses a big corpus as a database. Therefore, the main aim of this dissertation is to measure the gradient senses of the preposition in in an objective way, thus mapping the highly systematic semantic extension. Based on these gradient senses, the semantic and syntactic features of the preposition in produced by advanced Chinese English-major learners are investigated, combining quantitative and qualitative research methods. A quantitative analysis of the literal and other ten metaphorical senses of the preposition in is made at first. In accounting for the five factors influencing image schemata of each sense: “scale of Landmark”, “visibility”, “path”, “inclusion” and “boundary”, the formula of measuring the gradability of metaphorical degree is deduced: Metaphoricity=[[#Visibility] +[#Path] +[#Inclusion] +[#Boundary]]*[#Scale of Landmark]. The result is that the primary sense has the highest value:12, and all other extended senses have values down to zero. The more shared features with proto-scene, the higher the value of the metaphorical sense, and the less metaphorical the sense. EVENT and PERSON are the “least metaphoric” (value = 9-11); SITUATION, NUMBER, CONTENT and FIELD are “weak metaphoric” (value = 6-8); Also included are SEGMENTATION, TIME and MANNER (value = 3-5), and they are “strong metaphoric”; PURPOSE shares the least feature with proto-scene, and it has the lowest value, so it is “most metaphoric” (value = 0-2). Then, a corpus-based approach is employed, which offers a model for employing a corpus-based approach in Cognitive Linguistics. It compares two compiled sub-corpora: Chinese Master Academic Writing Corpus and Chinese Doctorate Academic Writing Corpus. The findings show that, on the semantic level, Chinese English-major students overuse in with a low level of metaphoricity, even advanced learners use the most metaphorical in rarely. In terms of syntactic behaviours, the most frequent nouns in [in+noun] construction are weakly metaphoric, whilst the nouns in the construction [in the noun of] are EVENT sense, which is least metaphorical. Moreover, action verbs tend to be used in the construction [verb+in] and [in doing sth.] in both master and doctorate groups. In the qualitative study, the divergent usages of the preposition in are explored. The preposition in is often substituted with other prepositions, such as on and at. The fundamental reason for the Chinese learners’ weakness is the negative transfer from their mother tongue (Wang 2001; Gong 2007; Zhang 2010). Although in and its Chinese equivalence zai...li (在...里) share the same proto-scene, there are discrepancies: the metaphorical senses of the preposition in are TIME, PURPOSE, NUMBER, CONTENT, FIELD, EVENT, SITUATION, SEGMENTATION, MANNER, PERSON, while those of zai...li (在...里) are only five: TIME, CONTENT, EVENT, SITUATION and PERSON. Thus the image schemata of each sense cannot be correspondingly mapped onto each other in different languages. This study also provides evidence for the universality and variation of spatial metaphors on the ground of cultural models. Philosophically, it supports the standpoint of Embodiment philosophy that abstract concepts are constructed on the basis of spatial metaphors that are grounded in the physical and cultural experience.

Many Chinese classifiers are nearly synonymic – they can be used with the same head nouns without changing the meaning of the sentence, in other words, such classifiers can be used interchangeably or almost interchangeably. This poses a challenge for Chinese language learners, especially those who lack such a grammatical category in their own native language. Another complication arises from the ambiguous English translations of many classifiers. In this paper we investigate the collocation behavior of near-synonymous Chinese classifiers, focusing on their semantic nuances and interchangeability. Analyzing 6 pairs of classifiers — 栋 and 幢, 匹 and 头, 批 and 派, 颗 and 粒, 辆 and 台, and 根 and 支— drawn from the HSK exam glossary, the dataset for this study encompasses 1200 samples (100 per each variable) and 416 distinct head nouns. Through a corpus-based approach we analyze collocation behavior of each classifier on its own and as a part of the pair. The results showcase that not all pairs exhibit complete interchangeability. The collocation behavior of 批 and 派 differ significantly, where 批 primarily quantifies batches with a 'first' connotation, while 派 is used more in artistic expressions. The interchangeability of 栋 and 幢 varies with context. 幢 emerges as the least fre¬¬quent morpheme in the corpus, emphasizing its specific contextual usage. While both are used in address lines, 栋 predominantly quantifies standalone buildings, whereas 幢 is more aligned with larger architectural complexes. The analysis of 匹 and 头 highlights their distinctiveness, with 匹 counting horses and wolves and 头 being more versatile with various animals. 颗 and 粒 appear partially interchangeable, particularly with 珠-related head nouns and items associated with plants, fruits, and trees. The research also underscores that 辆 is primarily linked to car-related nouns, while 台 is used more versatile as a classifier for machines and electronic devices, including computers, printers, phones, cameras. 根 and 支 only overlap in the head noun 笔, and their roles diverge, with 根 being a versatile classifier and 支 also appearing as part of medical terms.

This study compares simultaneously interpreted German speech to non-interpreted German discourse in order to determine whether interpreted language is characterised by any of the laws that have been found to feature in translated text, i.e. the law of growing standardisation and the law of interference. It is hypothesised that interpreters typically exaggerate German communicative norms, thereby producing manifestations of growing standardisation. In order to test this hypothesis, comparative and parallel analyses are carried out using corpora of interpreted and non-interpreted discourse. During the comparative phase, two types of interpreted German speech are each compared to non-interpreted language and to each other in order to determine how interpreted speech differs from non-interpreted discourse. During the parallel analysis, the interpreted German segments are compared to their source language counterparts with the aim of determining the reasons for the production of the patterns discovered during the first phase. The results indicate that interpreters do not produce patterns similar to those that characterise translated text: neither the law of growing standardisation nor the law of interference is manifest in the data. Instead, a different feature, namely an increased degree of generalisation, is discovered in the interpreters‟ output. This feature appears to be the result of the use of strategies that enable interpreters to deal with time, memory and linearity constraints inherent in SI. It can hence be confirmed that interpreted German differs from non-interpreted German discourse in certain respects. / Linguistics / M.A. (Linguistics)

The study adopts a theory of Descriptive Translation Studies to undertake a comparative analysis of stylistic devices in Shakespeare’s two plays, Julius Caesar and Macbeth and their Xitsonga translations. It contextualises its research aim and objectives after outlining a sequential account of theory development in the discipline of translation; and arrives at the desired and suitable tools for data collection and analysis.Through textual observation and notes of reading, the current study argues that researchers and scholars in the discipline converge when it comes to a dire need for translation strategies, but diverge in their classification and particular application for convenience in translating and translation. This study maintains that the translation strategies should be grouped into explicitation, normalisation and simplification, where each is assigned specific translation procedures. The study demonstrates that explicitation and normalisation translation strategies are best suited in dealing with translation constraints at a microtextual level. The sampled excerpts from both plays were examined on the preference for the analytical framework based on subjective sameness within a Skopos theory. The current study acknowledges that there is no single way of translating a play from one culture to the other. It also acknowledges that there appears to be no way the translator can refrain from the influence of the source text, as an inherent cultural feature that makes it unique. With no sure way of managing stylistic devices as translation constraints, translation as a problem-solving process requires creativity, a demonstration of mastery of language and style of the author of the source text, as well as a power drive characterised by the aspects of interlingual psychological balance of power and knowledge power. These aspects will help the translator to manage whatever translation brief(s) better, and arrive at a product that is accessible, accurate and acceptable to the target readership. They will also ensure that the translator maintains a balance between the two languages in contact, in order to guard against domination of one language over the other. The current study concludes that the Skopos theory has a larger influence in dealing with anticipating the context of the target readership as a factor that can introduce high risk when assessing the communicability conditions for the translated message. Contrariwise, when dealing with stylistic devices and employ literal translation as a translation procedure to simplification, the translator only aims at simplifying the language and making it accessible for the sake of ‘accessibility’ as it remains a product with communicative inadequacies. The study also concludes by maintaining that translation is not only transcoding, but the activity that calls for the translator’s creativity in order to identify and analyse the constraints encountered and decide on the corresponding translation strategies. / African Languages / D. Litt. et Phil. (African Languages)

This thesis aims to elucidate the translation process by devising a way of retrieving evidence of this process from its output. It further aims to assess the claims made by some scholars concerning the possible existence of Translation Universals. In order to isolate the interaction of texts and contexts, a corpus of American short stories was created, with their translations into Persian published after the 1979 Revolution. Three complementary methodologies gave a rounded picture: (1) Corpus-based Descriptive Translation Studies; (2) The pragmatic and rhetorically-based approach of Thinking Translation devised at St Andrews; and ‎(3) The analytical framework mostly established by Halliday in his Systemic Functional Grammar.‎ Approaching the process of translation in the specific order devised in this thesis provided four vantage points to analyse the data in a systematic way from linguistic, discourse, cultural and literary views before reaching what are at once the most personal and most characteristic aspects of a translator's work. The research begins with a literature review of the field and an account of linguistic constraints and of all Translation Universals hypothesised so far, followed by an extensive analysis of data in two consecutive chapters. With reference to the choices made in this corpus, it is discussed in the Conclusions chapter that most of the Translation Universals so far claimed are not in fact universal. It is the role of the translator which has emerged as the determining factor in producing a translated text, and thus as the key to resolving the issues explored in this thesis. It seems there are no constraints beyond the translator's reach, and there are no parameters which do not involve the translator, who introduces his or her own choices, or manipulates certain parameters. Only when they have done so, will the translation, as both process and product, be accomplished.

No Brasil, a maioria absoluta das primíparas, deseja parto normal logo que engravida, porém, em mais da metade dos casos, os nascimentos são cirúrgicos. O fenômeno da falta de correspondência entre o desejado e o efetivamente alcançado não é exclusividade das mulheres brasileiras, mas ocorre em vários países do ocidente. Por meio da Análise Crítica do Discurso (ACD) de relatos de parto normal após cesárea (relatos de VBAC, do inglês, vaginal birth after c-section) à luz da Linguística de Corpus (LC), buscamos elucidar o problema social entre a falta de correspondência entre o tipo de experiência desejada e a experiência obtida. O discurso dos relatos de VBAC nos parece ser o discurso ideal para desvelar os elementos dessa falta de correspondência, pois abordam tanto a experiência da cesárea anterior indesejada e, em regra, mal indicada, como a do parto desejado e alcançado. O recorte teórico-metodológico adotado reúne a ACD (Fairclough, 1989, 1992; Chouliaraki & Fairclough, 1999; Fairclough, 2003); a LC (Stubbs, 1993, McEnery & Wilson, 1997 e 2003, Tognini-Bonelli, 2001) e a Análise Crítica do Discurso Baseada em Corpus (Baker et al 2008; Baker, 2013; Baker & McEnery, 2005; Flowerdew, 2014). Para o estudo, foi compilado um corpus eletrônico em inglês e português. O corpus é composto por textos escritos pelas mulheres que passaram pela experiência de VBAC e não inclui textos mediados (entrevistas e relatos escritos por terceiros não foram incluídos). O Corpus BRABA (Corpus eletrônico de relatos de parto de mulheres brasileiras, estadunidenses, britânicas e australianas) se divide, respectivamente, em quatro subcorpora: Corpus BRA (93 relatos, 250 807 palavras), Corpus EUA (101 relatos, 225 736 palavras), Corpus UK (97 relatos, 92 197 palavras) e Corpus AU (92 relatos, 200 639 palavras. Os primeiros dois subcorpora Corpus BRA e Corpus EUA foram selecionados para esta pesquisa que pretende investigar como as identidades e a experiência do nascimento são representadas nos relatos de mulheres brasileiras e americanas e por meio dessa investigação chegar a elementos que elucidem o problema social. O processamento eletrônico valeu-se do programa AntConc 3.4.0w (Anthony, 2012) e das ferramentas da LC (listas de frequência, lista de palavras-chave, linhas de concordância, padrões lexicais, etc.). A análise foi direcionada pelas palavras-chave que correspondem aos sujeitos envolvidos e pelos colocados mais estatisticamente relevantes dessas palavras. No Corpus BRA, foram analisadas: eu (colocados: desisto, renasci, mamava); bebê (encaixado, morrer/morresse, sexo, batimentos, alto); marido (companheiro, apoiou, cortou); doula (amada, obstetriz, querida, presença); médico (fofa/fofinha, mudar/mudei, cesarista, ginecologista, humanizada); anestesista; enfermeira (obstétrica/obstetra, cadê, soro, chamar); parteira (liguei/ ligar, doula, casa); obstetriz (doula, toque). No Corpus EUA: I (wish, protested, lamented); baby (pound, girl, boy); midwife (certified, asst/assistant, student, assist); doula (hired, friend, called); nurse (practitioner, tells, triage); doctor (office, seen, comes); anesthesiologist; husband (poor, run, children). A análise permitiu que fosse elucidado o problema social em ambas as sociedades e fossem reveladas diferenças discursivas e culturais. A falta de correspondência entre a experiência desejada e a alcançada é representada como tendo sido causada pela sucessão de eventos distintos. Contudo, em ambos os corpora, as experiências são representadas e a autoidentidade e as identidades construídas discursivamente sob a égide dos traços da modernidade, marcadamente, em relação à reflexividade exercida ideologicamente. Porém, a reflexividade é operada, não só como um modo de sustentar, mas principalmente como forma de transformar as relações de dominação. / In Brazil the vast majority of primiparous women, on discovering that they are pregnant, hope to have normal deliveries. However, in over half of such cases surgical deliveries ensue. This mismatch between what pregnant women desire and what they actually experience is not exclusive to Brazil, but takes place in several Western countries. Through Corpus Linguistics (CL)-based Critical Discourse Analysis (CDA) of vaginal birth after c-section (VBAC) stories we seek to shed light on the social problem of a mismatch between the desired experience and the actual experience. VBAC stories seemed to us the ideal discourse for revealing elements of this mismatch, since they address both the experience of an unwanted (and usually wrongly indicated) prior C-section and that of the desired, and achieved, delivery. The theoretical-methodological approach we have adopted brings together CDA (Fairclough, 1989, 1992; Chouliaraki & Fairclough, 1999; Fairclough, 2003); CL (Stubbs, 1993, McEnery & Wilson, 1997 e 2003, Tognini-Bonelli, 2001), and Corpus-based Critical Discourse Analysis (Baker et al 2008; Baker, 2013; Baker & McEnery, 2005; Flowerdew, 2014). An electronic corpus was compiled in English and Portuguese for this study. The corpus is made up of texts written by women who have experienced VBAC and includes no mediated texts (i.e. interviews and third-party reports). The BRABA Corpus (Corpus of the birth stories of Brazilian, American, British and Australian women) encompasses four subcorpora respectively: Corpus BRA (93 stories, 250,807 words), Corpus USA (101 stories, 225,736 words), Corpus UK (97 stories, 92,197 words), and Corpus AU (92 stories, 200,639 words. The first two of these subcorporaCorpus BRA and Corpus USAwere chosen for this study, which investigates how identities and birth experiences are represented in the accounts of Brazilian and American women, and thus through this investigation uncovers elements that will shed light on the selected social problem. The computer processing used AntConc 3.4.0w (Anthony, 2012) and CL tools (frequency lists, keyword lists, concordance lines, etc.). Analysis was guided by keywords corresponding to the people mentioned in the stories and by the most statistically significant collocates of these keywords. From Corpus BRA the words were: eu (collocates: desisto, renasci, mamava); bebê (encaixado, morrer/morresse, sexo, batimentos, alto); marido (companheiro, apoiou, cortou); doula (amada, obstetriz, querida, presença); médico (fofa/fofinha, mudar/mudei, cesarista, ginecologista, humanizada); anestesista; enfermeira (obstétrica/obstetra, cadê, soro, chamar); parteira (liguei/ligar, doula, casa); obstetriz (doula, toque). From Corpus USA: I (wish, protested, lamented); baby (pound, girl, boy); midwife (certified, asst/assistant, student, assist); doula (hired, friend, called); nurse (practitioner, tells, triage); doctor (office, seen, comes); anesthesiologist; husband (poor, run, children). Analysis enabled this social problem to be laid bare in both societies, revealing discourse and cultural similarities and differences. The mismatch between the desired and the experienced outcomes is represented as having been caused by a succession of discrete events. In both corpora, experiences are represented, and self-identity and other identities are notably constructed in discourse under the aegis of features of modernity, above all, under reflexivity, which, in the discourses of VBAC stories takes place through empowerment, understood as self-actualization through newly gathered knowledge and ensuing courses of action/measures (Giddens, 2002).

A presente pesquisa tem como objetivo analisar, do ponto de vista descritivo, como foram realizadas as traduções de marcadores culturais presentes em duas obras da escritora afro-americana Toni Morrison: The Bluest Eye, traduzida por Manoel Paulo Ferreira como O Olho Mais Azul, e Beloved, traduzida por José Rubens Siqueira como Amada. Pretende-se analisar as escolhas e tendências tradutórias de cada tradutor diante das diferenças culturais entre a cultura fonte e a cultura meta. Para a realização deste estudo, apoiamo-nos no arcabouço teórico-metodológico dos Estudos da Tradução Baseados em Corpus (BAKER, 1993, 1995, 1996, 2000; CAMARGO 2005, 2007) e da Linguística de Corpus (BERBER SARDINHA, 2004).Para a investigação de marcadores culturais nos baseamos na reformulação realizada por Aubert (1981, 2006) a partir do trabalho sobre domínios culturais de Nida (1945). Para a extração e análise dos termos utilizamos o software WordSmith Tools versão 7.0. O presente estudo contém análises descritivas de 14 marcadores culturais selecionados a partir da lista de palavraschave gerada pelo WordSmith Tools, descrevendo as tendências e padrões tradutórios presentes nos textos meta e considerando, no âmbito da tradução, a especificidade dos romances e as diferenças culturais e históricas entre os Estados Unidos e o Brasil nos contextos históricos dos romances. / In this research, we intend to analyse, from the descriptive point of view, how the translations of cultural markers were held in two novels written by the Afro-American writer Toni Morrison: The Bluest Eye, translated by Manoel Paulo Ferreira as O Olho Mais Azul, and Beloved, translated by José Rubens Siqueira as Amada. We intend to analyse the choices and the translation tendencies of each translator before the cultural differences between the source culture and target culture. Our theoretical basis for the development of this research study is the theoretical and methodological approach of Corpus-Based Translation Studies (BAKER, 1993, 1995, 1996, 2000; CAMARGO 2005, 2007) and Corpus Linguistics (BERBER SARDINHA, 2004). We are also based on the study of cultural domains developed by Nida (1945) and reformulated by Aubert (1981, 2006). For term extraction and analysis, we used the WordSmith Tools software, version 7.0. So, this study contains descriptive analyses of 14 cultural markers selected from the keywords list created by WordSmith Tools, describing translation tendencies and patterns in the target texts and considering, in the scope of translation, the specificity of the novels and the cultural and historical differences between the United States and Brazil in the historical contexts of the novels.