Automated classification of texts into genres can benefit NLP applications, in that the structure, location and even interpretation of information within a text are dictated by its genre. Cross-lingual methods promise such benefits to languages which lack genre-annotated training data. While there has been work on genre classification for over two decades, none has considered cross-lingual methods before the start of this project. My research aims to fill this gap. It follows previous approaches to monolingual genre classification that exploit simple, low-level text features, many of which can be extracted in different languages and have similar functions. This contrasts with work on cross-lingual topic or sentiment classification of texts that typically use word frequencies as features. These have been shown to have limited use when it comes to genres. Many such methods also assume cross-lingual resources, such as machine translation, which limits the range of their application. A selection of these approaches are used as baselines in my experiments. I report the results of two semi-supervised methods for exploiting genre-labelled source language texts and unlabelled target language texts. The first is a relatively simple algorithm that bridges the language gap by exploiting cross-lingual features and then iteratively re-trains a classification model on previously predicted target texts. My results show that this approach works well where only few cross-lingual resources are available and texts are to be classified into broad genre categories. It is also shown that further improvements can be achieved through multi-lingual training or cross-lingual feature selection if genre-annotated texts are available in several source languages. The second is a variant of the label propagation algorithm. This graph-based classifier learns genre-specific feature set weights from both source and target language texts and uses them to adjust the propagation channels for each text. This allows further feature sets to be added as additional resources, such as Part of Speech taggers, become available. While the method performs well even with basic text features, it is shown to benefit from additional feature sets. Results also indicate that it handles fine-grained genre classes better than the iterative re-labelling method.

This thesis posits Enlightenment articulations of sympathy, in its capacity for establishing connections and its failures, as an appropriate methodology to articulate transatlantic literary exchange. Focusing on the sympathetic gap, the space sympathy must traverse, this thesis investigates the effect of emigration and place on genre and follows the trajectory from documentary to fictive forms and from a small gap to one unable to be bridged. Because the gap of sympathy is a spatial argument, the distance between is crucial as it indicates relationship. The introduction outlines my argument, with particular attention to transatlantic criticism, what is meant by the gap of sympathy, and the triad of place, emigration and genre. The first chapter discusses how Adam Smith articulated how one person is able to maintain a stable identity and is able to connect with another through imaginative comparison. The chapter establishes the trajectory of sympathy as the gap moves from smallest to unbridgeable, through comparison, sympathy and the failure of sympathy. In a series of case studies, Chapters Two through Five test out Smith’s theories in literary works; they examine the trajectory of transatlantic sympathy, where the gap moves from rhetorically being small to gaping, and moves generically from documentary forms to fiction. Chapter Two uses emigration guides written by British emigrants, who, because of their emigrant status, write from both an American and British perspective. The guides, because of their promotional intent, tend to underplay the gap of sympathy. Although they could be read as documentary and objective, the guides evidence ideological and rhetorical similarities to transatlantic fiction and thus serve as an entrance into the themes and stylistics one tends to associate with literary genres. Chapter Three examines the transatlantic correspondence of the Kerr family. As the Kerr family corresponds transatlantically (separated in space by the Atlantic and in time by more than 50 years), the issue of space becomes paramount to understanding the correspondence as well as if sympathy works in this generic register. Generically, the transatlantic letter is meant to provide virtual presence amid long stretches of absence; it also becomes an analogue for the absent other and the means by which the family may continue to be imagined across the gap of sympathy. Chapter Four examines Susanna Rowson’s transatlantic works, particularly Charlotte Temple, Slaves in Algiers, and Reuben and Rachel. Rowson’s own emigrant experience provides an entrée to the pain of transcultural sympathy that we see most clearly in Reuben and Rachel. Throughout her works Rowson also advocates a sympathy that is active and moral, rather than emotionally vacuous. Reuben and Rachel illustrates the gap of sympathy being bridged most effectively in cross-cultural adaptations and yet finally settles for a sympathy that must acknowledge separation and difference as well. Chapter Five explores the failures of sympathy and sociability present in Charles Brockden Brown’s gothic novels, Wieland and Edgar Huntly. Characters’ frontier locations and claustrophobic versions of sociability, as well generically, the gothic turn and failure of epistolary exchange, signals the moral ambiguity connected with becoming ‘this new man’ of America. Brown’s epistolary fiction briefly considered offers another generic attempt to examine how the gap of sympathy may be bridged and extend beyond the confines of the family. The Afterword points to the total breakdown of sympathy as a turn inward and away from sociability, where the self becomes frantic and frenetic (as evidenced by Crèvecoeur’s Letters from an American Farmer); it points to some useful applications to the gap of sympathy for transatlantic literary studies.

Notre propos consiste à décrire et à analyser certains témoignages recueillis par la Fondation Auschwitz en nous inspirant des recherches menées, notamment, en sociolinguistique, en ethnographie de la communication, en pragmatique, en Analyse des Conversations et en Analyse du Discours. Il s’agit, en d’autres termes, d’évaluer l’incidence que revêt, au niveau linguistique, l’inscription de la parole des témoins dans des matrices socio-discursives et interactionnelles particulières (discours oral, témoignage, entretien filmé, récit de l’expérience concentrationnaire). Pour cette étude, nous avons donc privilégié une approche transversale des données qui consiste à observer de manière circonstanciée, et à recenser statistiquement, un ensemble bien cerné de phénomènes microdiscursifs de surface : certains mots du discours (connecteurs et autres marqueurs discursifs) et certaines séquences discursives spécifiques (interventions des interviewers, gloses, séquences justificatives, etc.). Bien qu’elle se veuille principalement inductive et descriptive, notre démarche consistera à pratiquer un va-et-vient incessant entre une observation micro-analytique et une mise en perspective macro-analytique. En somme, partant de phénomènes très fins, nous aimerions décrire ce qui fait, globalement, la particularité de nos données, caractériser le comportement interactionnel et discursif de nos témoins et, ainsi, mettre en évidence certaines des régularités qui relient entre eux les divers objets de notre corpus. En d’autres mots, nous aimerions montrer comment l’examen des phénomènes microlinguistiques permet d’appréhender des mécanismes plus généraux. Les niveaux d’analyse micro- et macro-discursifs, loin de fonctionner de manière dichotomique comme c’est habituellement le cas, seront donc envisagés de manière interdépendante. Notre dissertation s’articule en cinq grandes parties, qui abordent toutes, peu ou prou, la notion de genre. Nous commençons (partie I) par envisager cette question de manière frontale, en essayant de définir le concept de genre en termes théoriques. Il s’agit d’expliquer clairement en quoi la problématique de l’appartenance générique peut s’avérer pertinente pour l’appréhension et la description de nos données. Dans la deuxième partie de ce travail, nous offrons une caractérisation très générale et succincte des récits sur la Shoah. Nous soulignons, notamment, les particularités nées du contexte politique et socioculturel où ces textes ont été produits. Ces précisions nous permettent de présenter plus en détail nos propres données, et de les comparer aux autres témoignages relatifs au phénomène concentrationnaire. Entrant ensuite dans le vif du sujet, nous consacrons notre troisième partie à la dimension testimoniale des données. Dans un premier temps, nous essayons de définir la notion même de témoignage, en cernant le plus précisément possible les contraintes définitoires propres à l’énonciation testimoniale et les questions épistémologiques qu’elles peuvent engendrer. Dans un second temps, et en accord avec notre projet méthodologique, nous relevons et analysons un certain nombre d’indices linguistiques qui nous paraissent spécifiques du discours testimonial. Dans la quatrième partie, qui se veut plus personnelle, nous nous attardons sur la dimension interactionnelle de notre corpus. Nous commençons par décrire les aspects liés à l’organisation structuro-séquentielle des interactions, en présentant deux modèles d’analyse élaborés à cet effet : le système de prise de tour et le modèle des paires adjacentes. Nous décrivons ensuite, dans ce cadre, les traits caractéristiques propres au format de l’interview. Enfin, grâce à l’analyse – qualitative et quantitative – à la fois des interventions produites par les interviewers et de certains marqueurs microdiscursifs utilisés par les interviewés, nous nous efforçons non seulement de cerner les phénomènes microdiscursifs propres au mode interactionnel de l’entretien, mais aussi de souligner, de façon peut-être paradoxale, les transgressions qui s’observent par rapport à ce format prédéterminé. Ce faisant, nous rejoignons nos réflexions antérieures sur l’instabilité inhérente à la notion de genre. Pour boucler la boucle, nous revenons sur cette question dans notre cinquième partie, où nous illustrons les réussites et les échecs discursifs de nos témoins par l’analyse détaillée de trois extraits qui mettent en évidence les défis rhétoriques liés à la complexité générique des discours envisagés.

No description available.

No description available.

The aim of this study was to examine ways in which High School Students approach academic writing. Both discursive and critical-analytical levels in 21 PM-texts and 14 investigating literary essays were analyzed through a genre theoretical and sociocultural framework in order to widen the understanding of students’ usage of explicit textual and interpersonal markers as well as their techniques of writing abstracts. The results showed that nearly 25 percent of the putative abstracts consisted of sampled and reshaped text sequences. The majority of these were on the other hand results of a creative process in which the students reshaped utterances from different sources. The results also showed that the essays in some aspects were subjective reviews rather than objective academic texts, but also constituted by both strategic and academic language actions. The PM-texts were in some ways products of the "old" school genre argumentative texts where explicit metatextual commentaries and language actions were relatively absent. The ways in which the students approach academic writing is therefore depending on how they perceive different school genres.

This study seeks to determine whether grammatical and presentational features of authentic non-fiction French texts are in any way related to the level of difficulty of texts as perceived by students who have taken Advanced Subsidiary Level (AS). The notions of text and genre are examined, as are the processes of reading in the first language (L1) and in the foreign language (L2). The question of ‘readability, and that of ‘authenticity’ in L2 are also examined. In order to ascertain students’ reactions to different text-types, 150 texts from French sources were gathered and classified. 100-word samples of each text were analysed for various linguistic features. Statistical tests on these were carried out, as well as statistical tests on the visual elements and layout of the whole texts. Further linguistic analysis was carried out within the text-type groups in order to ascertain their grammatical features. 31 students from local tertiary institutions were interviewed, and their perceptions on a sample of the texts were sought. This included the grading of texts for difficulty on a 1-5 Likert scale. The results of the interviews were triangulated with the statistical and linguistic analyses. A relationship was found between text-type and level of perceived difficulty. In the light of these results, the distinction between genre and text-type was examined, and a way was found of linking these into a textual taxonomy, which has close relations with the grammatical and presentational features which characterise the various text-types. In the light of these results, the question of the ‘topic approach’ to the teaching of French is examined, and a way found by which text-types that are perceived to be simpler are studied before those which are perceived to be more difficult. This approach advocates a more logical continuum of grammatical presentation than has hitherto been witnessed in course books for French at this level, while retaining the semantic integrity inherent in the ‘topic approach’.

Denna studie har i syfte att språkvetenskapligt undersöka hur reklamgenren förädras i ett komprimerat format, formatet googleannonser. För att få fram de språkliga dragen analyseras 30 st googleannonser ur ett genreperspektiv med innehålls-, strukturell och interpersonell analys som metod. Resultatet visar på ett flertal gemensamma drag med traditionell reklamgenre men uppvisar också skillnader. En strukturell skillnad är att googleannonsens rubrikdel tenderar att vara hybrid, de innehåller ofta både information om tjänster och produkter tillsammans med företagsnamn. Andra språkliga drag från materialet är bortprioriteringen av innehållsliga och interpersonella drag annars vanligt återkommande i traditionell reklamgenre. De gemensamma dragen för googleannonserna skulle kunna ingå i en ny modifierad subgenre till traditionell reklam, där nya konventioner för utformingen utvecklats.  The purpose of this study is to explore how the advertising genre adapts to a compressed format, from a linguistic perspective. The format being Google ads. 30 Google ads are analyzed with a content, structural and interpersonal method to distinguish linguistic features. The result shows multiple linguistic features related to traditional advertising genre. The study also finds diffrences. One divergent structural feature is the Google ads headline, who tends to have a hybrid function that includes both information about the product and information about the company. The material differ in comparison to traditional advertinsing genre in content and structure features, where some parts have been prioritized due to the limiting format. The new features for the Google advertising genre could be a part of a new modified subgenre, where new conventions has developed.

Denna undersökning är en kvantitativ analys av elevtexters ordklassfördelning och ordklassers semantiska egenskaper.Det här arbetet ämnar undersöka eventuella samband mellan fördelningen av ordklasser och deras semantiska egenskaper i elevtexter av berättande karaktär och de olika betyg som de givna texterna erhåller.Dessutom undersöks dessa grammatiska variablers funktion för skrivandet i ett försök att lyfta och granska grammatikundervisningens roll inom den svenska skolan – och vidare diskutera och motivera hur grammatikundervisningen kan utformas för att höja funktionaliteten och relevansen, och därigenom förhoppningsvis kunna öka intresset för grammatik hos eleverna. Vi ämnar även utreda och ge en överblick över vad inom ordklasserna och deras semantik som kan vara värdefull kunskap för utformningen av undervisningen och för elevernas textskapande.Urvalet består av autentiska elevtexter hämtade ur nationella proven i svenska skrivna av årskurs nio läsåret 2013. Texterna har analyserats manuellt med ett kvantitativt stilistiskt angreppssätt.Resultatet påvisar några generella samband mellan en del olika variabler och de olika betygsnivåerna. En högre användningsfrekvens av adjektiv tenderar att ge ett högre betyg, så även för abstrakta substantiv. Högre betyg kan också förknippas med en högre användning av mentala och relationella processer – och därigenom har texterna på A-nivå även en högre andel statiska verb, då dessa ofta realiseras genom relationella processer. Texterna på A-nivå uppvisar också ett enhetligare tempusbruk jämfört med texterna på E-nivå, som istället tenderar att mer frekvent skifta tempus på ett oregelbundet sätt. Texterna på E-nivå upprepar gärna också samma ord i större utsträckning än de övriga betygsnivåerna. Undersökningen och resultatet kopplat till litteraturen och tidigare forskning motiverar en genrebaserad undervisning som ökar elevernas medvetenhet om olika texters uppbyggnad.

Mémoire numérisé par la Division de la gestion de documents et des archives de l'Université de Montréal.