No description available.

Many texts we encounter in our everyday lives are lexically and syntactically very complex. This makes them difficult to understand for people with intellectual or reading impairments, and difficult for various natural language processing systems to process. This motivated the need for text simplification (TS) which transforms texts into their simpler variants. Given that this is still a relatively new research area, many challenges are still remaining. The focus of this thesis is on better understanding the current problems in automatic text simplification (ATS) and proposing new data-driven approaches to solving them. We propose methods for learning sentence splitting and deletion decisions, built upon parallel corpora of original and manually simplified Spanish texts, which outperform the existing similar systems. Our experiments in adaptation of those methods to different text genres and target populations report promising results, thus offering one possible solution for dealing with the scarcity of parallel corpora for text simplification aimed at specific target populations, which is currently one of the main issues in ATS. The results of our extensive analysis of the phrase-based statistical machine translation (PB-SMT) approach to ATS reject the widespread assumption that the success of that approach largely depends on the size of the training and development datasets. They indicate more influential factors for the success of the PB-SMT approach to ATS, and reveal some important differences between cross-lingual MT and the monolingual v MT used in ATS. Our event-based system for simplifying news stories in English (EventSimplify) overcomes some of the main problems in ATS. It does not require a large number of handcrafted simplification rules nor parallel data, and it performs significant content reduction. The automatic and human evaluations conducted show that it produces grammatical text and increases readability, preserving and simplifying relevant content and reducing irrelevant content. Finally, this thesis addresses another important issue in TS which is how to automatically evaluate the performance of TS systems given that access to the target users might be difficult. Our experiments indicate that existing readability metrics can successfully be used for this task when enriched with human evaluation of grammaticality and preservation of meaning.

This study investigates the Swedish translations of extended premodifiers in a non-fictional German text on pedagogical theory and practice. The aim is to analyse complications and implications connected to different translation options and strategies. The analysis is based on Solfjeld’s (2003, 2004) German-Norwegian translation studies, which propose a classification model for translation strategies for German extended premodifiers with regard to aspects of explicitation and implicitation. The results show relations between postnominal finite target clauses and explicitation on the one hand, and on the other hand between prenominal non-clausal units and implicitation. Explicitation tends to result in sentence splitting which affects the text length and text cohesion.In addition, the study defines three main categories of German extended modifiers. The first category of adverbial extender premodifiers describe time, degree, and perspective and can easily be transferred into Swedish premodifying structures. The second category includes objects or predicatives which must be transferred into postmodifying structures. For the extended premodifiers belonging to the third category, the transferring into pre- or postmodifying target structures is optional. The study shows that the complexity of the extended premodifiers is a crucial factor which determines the choice of translation strategy. While explicifying relative clauses have a high amount of correspondence with Swedish language structures, the analysis emphasizes that implicifying premodifying target structures have less negative impact on text length, sentence splitting and text cohesion.

This paper studies the translation of style in the book Die Welt auf dem Teller by Doris Dörrie (2020) from German into Swedish. It is argued that the style is partly held in punctuation and sentence length and the focus of the study is how these can be translated from German into Swedish, considering their importance for the text style.  The analysis shows that the source text has a more differentiated use of punctuation whereas the target text is more restricted/neutral. Differences concerning how punctuation is translated are mainly due to grammatical differences but also a question of whether the punctuation can be experienced equally by the source and target text reader. Both circumstances have an influence on the translation.  It is argued that Swedish readers expect shorter sentences and texts accessible to the reader. A higher density and more complex constructions are accepted in German. On average, shorter sentences are found in this translation, and the deviance is lower. When the sentence length in the source text is seen as a significant stylistic feature, this structure is kept in the translation. Otherwise, sentences are often split to become more accessible. When the sentence construction is kept, it is however still often shorter than the source text. As this is expected from the target text reader, the stylistic effect can arguably still be considered preserved.  The translation can therefore be said to be more neutral on both sentence length and punctuation. The translation is giving the stylistic features space when this is considered a characteristic feature, otherwise, it is changed to fit the language norm.