Sequence-to-sequence learning led to significant improvements to machine translation (MT) and automatic speech recognition (ASR) systems. These advancements were first reflected in spoken language translation (SLT) when using a cascade of (at least) ASR and MT with the new "neural" models, then by using sequence-to-sequence learning to directly translate the input audio speech into text in the target language. In this thesis we cover both approaches to the SLT task. First, we show the limits of NMT in terms of robustness to input errors when compared to the previous phrase-based state of the art. We then focus on the NMT component to achieve better translation quality with higher computational efficiency by using a network based on weakly-recurrent units. Our last work involving a cascade explores the effects on the NMT robustness when adding automatic transcripts to the training data. In order to move to the direct speech-to-text approach, we introduce MuST-C, the largest multilingual SLT corpus for training direct translation systems. MuST-C increases significantly the size of publicly available data for this task as well as their language coverage. With such availability of data, we adapted the Transformer architecture to the SLT task for its computational efficiency . Our adaptation, which we call S-Transformer, is meant to better model the audio input, and with it we set a new state of the art for MuST-C. Building on these positive results, we finally use S-Transformer with different data applications: i) one-to-many multilingual translation by training it on MuST-C; ii participation to the IWSLT 19 shared task with data augmentation; and iii) instance-based adaptation for using the training data at test time. The results in this thesis show a steady quality improvement in direct SLT. Our hope is that the presented resources and technological solutions will increase its adoption in the near future, so to make multilingual information access easier in a globalized world.

This thesis is a study of the inter-institutional translation process through which the Swedish versions of EU legislative acts are created, focusing on the conditions for plain language work within this process. These Swedish translations have long been considered incomprehensible. Complicated originals and strong demands for close correspondence to the source text have been considered the main reason. This study aims to examine the translation process, to see how institutional factors shape the scope for plain language work. The theoretical and methodological bases of the study are the sociology of translation, focusing on translation as a social activity, and ethnography, focusing on the participants’ perspective. The data analysed consist of field notes, interviews, video recordings of participants translating, and revised translations. A key factor for plain language is shown to be the demands for close correspondence to previous texts – not only the source text but also previous Swedish translations. Close correspondence is necessary due to special demands on legal translation but is also supported by other institutional factors, such as translators’ insecurity regarding legal language, short deadlines and the needs of internal readers. The use of CAT tools such as TWB further reinforces this dependence on previous texts. Other institutional factors that stand out are the relative isolation of the translation units, with limited feedback from outside the unit, and the position of the translated documents within the EU legal system. The results indicate that the difficulties for a Swedish reader are often not due to linguistic or textual formulation of the documents, but to the gap this EU position creates, between the reader’s (Swedish) and the text’s (EU) frame of reference respectively. Support for plain language is strong at the units, but the concept appears to have been partially given a local meaning, showing that the traditional understanding of plain language may not be fully applicable in this setting. This shows the necessity to identify the aspects most central to comprehensibility, for different types of texts and in different settings – a necessity not only for plain language work within the EU, but in national settings as well. / I avhandlingen studeras den interinstitutionella översättningsprocessen, till svenska, för EU:s rättsakter, med fokus på hur villkoren för klarspråksarbetet ser ut inom denna process. De svenska översättningarna har länge ansetts svårbegripliga. Svårt skrivna original och starka krav på nära överensstämmelse med källtexten har pekats ut som huvudsakliga skäl till detta. Undersökningen syftar till att ta reda på hur institutionella faktorer formar utrymmet för klarspråksarbete inom översättningsprocessen. De teoretiska och metodiska utgångspunkterna för undersökningen är översättningssociologi, som betraktar översättning som en social aktivitet, och etnografi, som betonar deltagarperspektivet. Materialet består av fältanteckningar, intervjuer, videofilmade observationer av översättnings­arbete samt granskade översättningar. Krav på nära överensstämmelse med tidigare texter visar sig vara centralt för klarspråksarbetet; inte bara överensstämmelse med källtexten utan också med tidigare svenska översättningar. Den nära överensstämmelsen är nödvändig till följd av de särskilda krav som ställs på juridisk översättning, men stöds också av andra institutionella faktorer, exempelvis översättarnas osäkerhet omkring juridiskt språk, korta tidsfrister och interna läsares behov. Användandet av CAT-verktyg som TWB förstärker ytterligare denna bundenhet till tidigare texter. Andra institutionella faktorer är enheternas relativa isolering, med begränsad återkoppling utifrån, och den position som de översatta dokumenten har inom EU:s juridiska system. Resultaten pekar på att svårigheterna för en svensk läsare ofta inte är kopplade till texternas språkliga eller textuella utformning, utan till det avstånd som skapas mellan läsarens (svenska) och textens (EU-) referensram, till följd av just textens position inom EU-systemet. Stödet för klarspråk är starkt på enheterna, men begreppet förefaller ha getts en delvis lokal innebörd, vilket visar att den traditionella innebörden av klarspråk kanske inte är helt tillämpbar i det studerade sammanhanget. Detta visar behovet av att identifiera de aspekter som är mest centrala för begripligheten, för skilda sorters texter och i skilda kontexter – en nödvändighet inte enbart för klarspråksarbetet inom EU, utan även i nationella sammanhang.

This master's thesis consists of a Czech language translation and commentary for selected chapters of Dans l'enfer des tournantes, an autobiography by Samira Bellil, a French author of Algerian descent. The commentary is divided into two parts: a presentation of literature written by French authors of Maghrebi descent, the so-called beurs, the language they use in literature and reception of their works; and an analysis of the source text and its reception, the problems encountered during the process of translation and their solutions as well as the chosen translation method. Key words: Samira Bellil, autobiography, beurs, translation, translation analysis, translation method, translation problem, translation shift, suburbs of Paris

No description available.

This paper considers some of the questions posed by literary translations both from and into Swahili. While the questions a translator might address as she proceeds with each translation may be the same, their differing answers often highlight the translator’s different position towards, and history with, each target language, as well as her aesthetic and political commitments in each. The projects discussed are Mlenge Fanuel Mgendi’s comic short story Starehe gharama (Comfort is Expensive) about a young schoolboy’s misadventure on a daladala bus in Dar es Salaam and Tope Folarin’s Caine Prize shortlisted story Genesis (Mwanzo), in which two Nigerian boys living in the American Midwest witness their mother’s struggle with her new surroundings.

No description available.

The aim of this Master Thesis is to present an annotated literary translation of two chapters of the novella Vstrecha from Maxim Osipov's book, Grech zhalovatsya. The first part of this Master Thesis contains a translation of selected chapters from the novella Vstrecha. The second part, a commentary on the translation, includes information about the author and his work, style classification, the role of translation, vocabulary, intertextuality, characterization of translation methods, strategy of solving translation problems and typology of translation shifts.

The aim of this Master Thesis is to provide the reader with an annotated literary translation of the book Palabras Envenenadas written by the Spanish author Maite Carranza. The first part of the thesis presents in total of 6 translated chapters of the novella Palabras Envenenadas. The second part focuses on presenting the information about the author and the style of the book as well as the role of the original and the translatated text in the source and target culture context, the development of the translation analysis and translation methods while discussing the translation problems and their solutions. Last but not least, the thesis should also offer the reader also a brief glimpse into a process of the birth of a translated literature. Powered by TCPDF (www.tcpdf.org)

La question directrice qui nous guide tout au long de ce travail est le "principium redendae rationis sufficientis" (le principe de la raison suffisante) sur la traduction duquel Heidegger mène ses réflexions au sujet de la traduction sans le mettre lui-même en relation avec la problématique de la traduction. En d’autres termes, la question que nous nous posons est de savoir si quelque chose comme la traduction est possible, quelle est la raison déterminante de cette possibilité et pourquoi en général une telle chose est possible plutôt qu’impossible. La question de la possibilité de la traduction au sens de ratio sufficiens, est une question qui porte sur l’essence de la traduction en tant que lieu où l’identité et la non-identité à la fois se thématisent et se problématisent. Nous nous référons afin de répondre à cette question directrice à la phénoménologie d’Edmund Husserl. Indépendamment du fait que Husserl lui-même, comme nous le montrons, pense à une telle possibilité de la traduction, l’intérêt de la phénoménologie de Husserl pour la théorie de la traduction consiste à proposer un instrument formidable d’analyse linguistique, dont la spécificité ne consiste pas dans l’analyse de la forme telle qu’elle est, mais dans l’analyse de la forme à partir du contenu de la conscience dans laquelle toute forme a soit son équivalent conscientiel conscient selon le mode intentionnel de la conscience, soit est explicable à partir d’un tel contenu. L’idée et la possibilité d’une recherche phénoménologique de la traduction se réalisent dans le présent travail à partir de l’exploration d’une telle conscience qui se présente en même temps comme une conscience traductionelle. / The central thought of the present work is the "principium redendae rationis sufficientis", the translation of which Heidegger himself dealt with in his reflections on translation, but without drawing these in relation to the issue of translation. In other words, we raise the following question: If something like translation is possible, what is then the decisive reason and why is translation more likely possible than impossible? The question with regard to the possibility of translation in the sense of ratio sufficiens enquires about the essence of translation as a specific location in which identity, but also non-identity, are simultaneously problematized and thematised. In order to elucidate our initial question, we refer to Edmund Husserl’s phenomenology. Irrespective of the fact that Husserl himself thought of a such possibility of translation, as we intend to show, the interest of Husserl’s phenomenology for a theory of translation consists in offering a formidable instrument of linguistic analysis whose specificity does not rely on a purely formal analysis, but on an analysis of form deriving from the content of consciousness, in which each form, according to the intentional mode of consciousness, is either disposing of its conscious equivalent or is explicable through a such content. The idea and the possibility of a phenomenological translation research is to carry out in this present work starting from the investigation of a such consciousness, that presents itself at the same time as a translation-consciousness. / Der dieser Arbeit zugrundeliegende Leitgedanke ist das "principium redendae rationis sufficientis", mit dessen Übersetzung sich Heidegger bereits in seinen Betrachtungen zu Übersetzung befasst hat, ohne jedoch diese selbst auf die eigentliche Übersetzungsproblematik zu beziehen. Mit anderen Worten fragen wir uns: Wenn etwas wie Übersetzung möglich ist, was ist der entscheidende Grund dafür und weshalb ist etwas derartiges vielmehr möglich als unmöglich? Die Frage nach der Möglichkeit der Übersetzung im Sinne der ratio sufficiens fragt nach dem Wesen der Übersetzung als Ort, an dem sich Identität und Nicht-Identität gleichzeitig thematisieren und problematisieren. Um unsere Ausgangsfrage zu erforschen, beziehen wir uns auf die Phänomenologie Edmund Husserls. Unabhängig davon, dass Husserl selbst, wie wir zeigen werden, bereits an eine solche Möglichkeit der Übersetzung gedacht hat, besteht das Interesse der Husserlschen Phänomenologie für die Übersetzungstheorie darin, ein ausgezeichnetes Instrument zur linguistischen Analyse vorzuschlagen, dessen Spezifizität nicht in einer reinen Formanalyse besteht, sondern in der Analyse der Form ausgehend vom Bewusstseinsinhalt, in dem jede Form entweder gemäß dem intentionalen Modus des Bewusstseins über ein bewusstseinsmäßiges Äquivalent verfügt oder von einem solchen Inhalt aus erklärbar ist. Die Idee und Möglichkeit einer phänomenologischen Übersetzungsforschung wird im Ausgang von der Erforschung eines solchen Bewusstseins, das sich auch als ein Übersetzungsbewusstsein darstellt, durchgeführt.

Jitka Nešporová doctoral thesis ABSTRACT The doctoral thesis focuses on the translation persona and work of Ludvík Kundera. A poet himself, he translated verse first and foremost and was able to do so from many languages, although he sometimes resorted to the use of interlinear translations. The research, however, concentrates primarily on the description, analysis and reception of his direct translations in the Czech - German language pair, which was predominant in Kundera's translation work. The thesis contributes to the understanding of the history of Czech literary translation after 1945 by describing Kundera's translation method, making accessible part of his literary estate, notably his correspondence, and providing an update to Kundera's translation bibliography. Kundera is best known as the exclusive, authorized translator of Bertolt Brecht's drama and poetry. His translations of Expressionists Georg Trakl and Gottfried Benn, for which he received the State Award for Literary Translation in 1996, are considered canonical today. The same holds true for his translations of Paul Celan's poetry and Alfred Kubin's novel Die andere Seite (The Other Side). Kundera was also the first to introduce Czech readers to the poetry of Alsatian Dadaist Hans Arp and the poetry of East German lyricist Peter Huchel....