Letteren en Wijsbegeerte

Departement Taalkunde

Doctoraatsverdedigingen

The situated construction of language ideologies in Aruba: a study among participants in the language planning and policy process - Eric Mijts (26/11/2021)

Eric Mijts

  • 26/11/2021: doctoraatsverdediging
  • Promotoren: Reinhild Vandekerckhove, Pol Cuvelier en Michael Meeuwis

Abstract

Dit onderzoeksproject gaat over de ingewikkelde relatie tussen taalideologie, kolonisatie en de sociaaleconomische mobiliteit en emancipatie in voormalige koloniën waar het overgeërfde koloniale taalbeleid doorklinkt in het taalbeleid van tegenwoordig. Dit taalbeleid geeft voorrang aan de taal van de voormalige kolonisator boven de thuistalen van de mensen in onderwijs, en ook vaak in de rechtspraktijk en in bestuur. Vooral daar waar het gaat over formele onderwijscontexten heeft deze koloniale erfenis een aantal knelpunten tot stand gebracht die grote invloed hebben op deze samenlevingen. Over het algemeen wordt de koloniale taal verkozen als onderwijstaal voor secundair en hoger onderwijs, en als de thuistaal al gebruikt wordt in basisonderwijs, dan is dat met als hoofddoel om de ontwikkeling van taalbeheersing in de koloniale taal te ondersteunen.

Ik heb aangetoond dat het handhaven van de koloniale taal in onderwijs en in andere domeinen van taalgebruik hand in hand gaat met ideologieën die de taal, cultuur, instituties en onderwijssystemen van de voormalige kolonisator als superieur voorstellen, terwijl de thuistalen worden gepresenteerd als inadequaat, ongeschikt voor onderwijs, of zelfs ook ongeschikt om als taal benoemd te worden. Aan de hand van juridische documenten, beleidsteksten, onderzoeksrapportages, gedrukte media en sociale mediadiscussies heb ik de persistentie van koloniale taalideologieën aangetoond doorheen de geschiedenis en doorheen verschillende tekstgenres in de meertalige kleine eilandstaat Aruba. Bovendien heb ik aan de hand van vijf case-studies aangetoond dat deze koloniale taalideologieën ook de basis zijn voor de hiërarchische positionering van thuistalen in andere landen. Deze structurele hiërarchische positionering wordt bewust en onbewust aanvaard en ondersteund door de meerderheid van de bevolking, terwijl juist weerstand bestaat tegen het inclusieve gebruik van de thuistalen voor onderwijs, rechtspraktijk en bestuur.

Ik trek de conclusie dat het begrip van ideologieën en geloofssystemen over taal, onderwijs en sociaaleconomische ontwikkeling de structurele en ideologische aard van ongelijkheid op basis van taalkeuze in meertalige dekoloniale staten kan blootleggen. Concreet maatschappelijk draagvlak voor transities die de talen van de meerderheid van de bevolking van voormalige kolonies bevorderen kan enkel tot stand komen als de persistente aannames over de minderwaardigheid van thuistalen tegenover de talen van de voormalige kolonisator geïdentificeerd, blootgelegd en kritisch geanalyseerd als overblijfselen van het koloniaal verleden. Pas als het structureel ideologisch karakter van het taalbeleidsdebat wordt erkend, kan het worden geadresseerd.

More Meaning Than Meets the Eye. Robust and Scalable Applications of Pre-trained Representations for Biomedical NLP - Pieter Fivez (27/09/2021)

Pieter Fivez

  • 27/09/2021: doctoraatsverdediging
  • Promotoren: Walter Daelemans & Simon Šuster 

Abstract

Pre-trained distributional representations of words and phrases have become omnipresent in naturallanguage processing (NLP), where they have led to significant improvements in machine learningperformance for a wide range of applications. Recent research has investigated to what extent theserepresentations are effective for tackling the challenges of the biomedical text domain. However, itremains difficult to properly disentangle the interplay of model architectures, training objectives,data sources, and downstream biomedical NLP tasks for which the representations are used as inputfeatures. As a result, it is still unclear to which extent these representations can be applied to encodespecific biomedical semantics for future applications which would require complex domainknowledge. 

In this thesis, we specifically explore what we consider to be robust and scalable applications ofpre-trained representations for biomedical NLP. These applications go against the current dominantparadigm in NLP research, which has achieved many successes by fine-tuning large and complexneural network architectures using vast amounts of data. In contrast, we explicitly try to minimizethe complexity of models that use the pre-trained representations, as well as the amount ofsupervised data necessary for developing the models, while keeping the models transferable acrossvarious domains and applicable in unsupervised ways, e.g. using distance metrics such as cosinesimilarity. 

While this paradigm can impose a performance ceiling on our proposed models compared to otherstate-of-the-art approaches, it also offers various benefits. Firstly, it helps to highlights thecontribution of various aspects of a method. For instance, it can emphasize the effectiveness oftraining objectives which work for models with low complexity. Secondly, it minimizes thecomputational cost of our proposed systems, and as such aims at contributing to more equitable anddemocratic NLP research. Lastly, the limitations of this paradigm also challenge us to explore novelapproaches that are more efficient. For example, we can compensate for less model complexity andtraining data by finding more effective training objectives.

Investigating Identity Work by Chinese Police Officers in Police Mediation Discourse - Wejing Feng (24/08/2021)

Wejing Feng

  • 24/08/2021: doctoraatsverdediging
  • Promotoren: Chen Xinren, Mieke Vandenbroucke en Jef Verschueren

Abstract

Identity is a hot topic for research in social sciences. However, little research explores identity in police mediation discourse. Meanwhile, existing research on identity from pragmatic perspective has mainly focused on identity construction and its linguistic representation, ignoring identity deconstruction and identity work as a means of rapport management. In response to this research gap, the present study, drawing on Pragmatic Identity Theory, Rapport Management Model, and Linguistic Adaptation Theory, and adopting a linguistic ethnographic methodology, aims to probe into the Chinese police officers’ identity work in their mediation interaction. Drawing on the transcripts of recordings of 13 Chinese police officer-mediated interactions in a Chinese context, this study examines what, how and why identity work is carried out by police officers in civil conflict mediation discourse. By providing a detailed description and interpretation of identity work in Chinese police mediation interaction, this study has the following implications:

  • Theoretically, this study can promote the understanding of identity work, particularly the characteristics and motivations of identity work in institutionally multiple-identity situation. The present study can enrich and broaden the scope of research on identity work and pragmatic identity by discussing in detail identity deconstruction and its linguistic devices, thus showing a more dynamic feature of identity work. Meanwhile, by bringing Rapport Management Model into the analysis of the communicative goals of identity work, this study provides a new perspective on the relationship between identity work and relational work. 
  • Practically, the present research can fill in the research gaps in both police discourse and mediation discourse researches by exploring the various identity (de)construction process in police mediation discourse, having shown the complexity and dynamics of police mediation discourse from the pragmatic perspective. It can enhance the present understanding of police mediation work, facilitate the civil conflict resolution and improve the effectiveness of police mediation work. 
  • Methodologically, the linguistic-ethnographic research method adopted in the present study could be used to examine identity work in other institutional settings. This method, by shadowing and observing the interactants’ discursive interactions, gives access to the original first-hand data, including natural recording, questionnaires, interviews, and field notes, etc. Therefore, this study may contribute to the study of identity work and police mediation discourse by providing multiple types of data and a different methodology for reference.

Computational approaches to intertextuality: from retrieval engines to statistical analysis - Enrique Manjavacas Arévalo (6/07/2021)

Enrique Manjavacas Arévalo

  • 06/07/2021: doctoraatsverdediging
  • Promotor: Mike Kestemont

Abstract

A common device exploited by literary writers consists in the reuse of texts originally authored by others. In doing this, literary writers–often unconsciously–establish references to other texts and, thus, situate their work in relation to others within a large network of texts. This device of literary referencing is not only responsible for a particular artistic experience–which is based on the recognition of the links–, but has also the potential effect of enriching the interpretation and the meaning of the work in a larger context. 

A famous example from classic western literature is Ovid's beginning to the Amores:

"arma gravi numero violentaque bella parabam" ("I was planning to write about arms and violent wars using a heavy meter'')

where Ovid reuses the beginning of Vergil's Aeneid:

"arma virumque cano Troiae qui primus ab oris'' ("I sing of arms and the man, who first [came] from the shores of Troy'')

In this case, the recognition of the parallelisms at multiple levels–e.g. the reuse of the word "arma'', the analogous syntactic construction using "-que'' (en. and), the adoption of the theme of war ("violenta bella'', en. violent wars) and the instruments of war ("arma'', en. arms), as well as the resemblance in metrical pattern–not only constitutes a core artistic value to this passage, but also let us situate the Amores–a love elegy–in relation to the epic literature of Vergil, to which it relates from the point of view of literary genres.

The study of these referential processes in literary works has been systematized through the theory of "intertextuality'', whose application can considerably profit from the identification of parallelisms at scale. In particular, the identification of biblical references in Medieval religious literature constitutes a task in which computational methods can make an impact, given the central role of the Bible in these writings.

This PhD thesis is concerned with the study of "intertextual'' links from a computational point of view, and aims at fulfilling two goals. The first one consists in improving the capacities of automated retrieval systems targeting cases of intertextual links. The second one emphasizes the usage of data-driven approaches to the study of intertextual processes. In order to accomplish these goals, this PhD thesis focuses on the use case of biblical references in the Latin Patrology–a large body of Medieval religious writings.

lt-cleft Change across the Lifespan in Early Modern Authors: Variation Rhetoric and the Individual in Constructional Change - William Standing (26/05/2021)

William Standing

  • 26/05/2021
  • Promotor: Peter Petré

Abstract

Het taalgebruik van het individu is over het algemeen onderbelicht binnen diachrone en syntactische studies, die zich voornamelijk richten op de details van taalverandering op het niveau van de gemeenschap. Wil men echter een volledig beeld krijgen van de aard van het taalgebruik in de loop van de tijd en van de cognitieve basis waaruit dit gebruik voortkomt, dan moet het individu centraal gesteld worden in het onderzoek naar de mechanismen van taalgebruik en verandering. 

Veel belangrijke eigenschappen van taalkundige systemen (zoals analogie, automatisering, gevolgtrekking en associatie) bestaan noodzakelijkerwijs in de eerste plaats in de hoofden van individuele sprekers, een feit dat stilzwijgend wordt voorondersteld in veel usage-based onderzoek, maar dat zelden wordt geoperationaliseerd. Dit geldt in het bijzonder voor de syntaxis, waarvan over het algemeen wordt aangenomen dat zij intergenerationeel en dus abrupt verandert. Het huidige onderzoek wil dit gebrek aan aandacht voor de individuele dynamiek van gebruik en verandering rechtzetten, en daarnaast ook de patronen die op gemeenschapsniveau worden waargenomen in kaart brengen, door een diepgaand kwalitatief en kwantitatief onderzoek aan de hand van de case van de gekloofde zinsconstructie die begint met it (de ‘it-cleft’, bv. It was Bill that kissed Mary). De it-cleft is een complexe biclausale zinsstructuur met een verscheidenheid aan betekenissen, waarbij de focusconstructie centraal staat. Door het onderzoeken van constructionele verandering in deze constructie aan de hand van een dataset van veelschrijvers, beoogt dit proefschrift vooruitgang te boeken in dit onderbestudeerde onderzoeksdomein. 

De beoogde doelen worden gerealiseerd binnen het theoretische kader van Construction Grammar, met integratie van sociolinguïstische analyse en kwantitatieve statistische methodologieën, en aan de hand van de analyse van 50 vroegmoderne Engelse auteurs geboren in de 17e eeuw. Speciale aandacht gaat daarbij uit naar de verandering die individuen hebben ondergaan tijdens de levensloop, de sociale dynamiek van de verschillende taalgemeenschappen, en de associaties die sprekers hebben gemaakt bij de verwerking en het gebruik van de it-cleft. Ook het schematiseringsproces van de gekloofde it-constructie gedurende de vroegmoderne periode wordt in detail beschreven met betrekking tot een aantal belangrijke vectoren van verandering, evenals een verheldering van de rol van individuen in dit proces.

Le conditionnel de reprise en français. Une étude de son origine et de son évolution de sens - Jessica Van de Weerd (21/04/2021)

Jessica Van de Weerd

  • 21/04/2021
  • Promotoren: Patrick Dendale en Anne Vanderheyden

Abstract

Sinds de jaren ’90 is de conditionnel een populair onderwerp in het linguïstisch onderzoek. Het is voornamelijk sinds de publicatie van het collectieve volume Le conditionnel en français van Dendale & Tasmowski in 2001 dat de interesse in dit onderwerp enorm gegroeid is. Het boek geeft een overzicht van de problemen, perspectieven en analyses van de studie van de conditionnel in het Frans. Er wordt ook een overzicht gegeven van de verschillende gebruiken.

Ons onderzoek gaat over de conditionnel épistémique 1 (vanaf nu CE1). De studie van dit specifiek gebruik van de conditionnel is de laatste tijd erg populair geworden. Hoewel het gebruik van de CE1 in synchrone linguïstiek uitgebreid bestudeerd is, is dit niet het geval voor diachrone linguïstiek. Met dit onderzoek willen wij dan ook een bijdrage leveren op dat laatste vlak. De weinige diachrone studies over de CE1 die er zijn kunnen we samenbrengen in drie thematieken:

(1°) de behandeling van de CE1 in Franse grammatica’s van de 16e tot de 20e eeuw

(2°) het verschijnen van de eerste voorbeelden van de CE1 in verschillende tekstgenres

(3°) de evolutie van de gebruiken van de conditionnel, waaronder ook de CE1

Om een antwoord te bieden op onze hoofdvraag – Wanneer en via welk mechanisme is de CE1 ontstaan? – gaan we systematisch in op elk van deze drie thematieken. Eerst bestuderen we de manier waarop Franse grammatica’s de CE1 semantisch beschrijven. We analyseren de verschillende termen en formuleringen die gebruikt worden bij de semantische beschrijving van de CE1. Deze studie geeft ons een globaal beeld van de betekeniselementen die, volgens de grammatica’s, een rol spelen bij de creatie van de semantische waarde van de CE1. Vervolgens analyseren we een corpus van oude voorbeelden van de CE1 uit juridische teksten. We verifiëren of de betekeniselementen die door de grammatica’s aangehaald worden ook echt te herkennen zijn in de oude voorbeelden van de CE1. Deze analyse laat ons toe om de co(n)texten waarin de oude CE1 voorkomt beter de definiëren. Tot slot formuleren we een hypothese over de oorsprong (en de evolutie) van de CE1. We verdedigen de hypothese volgens dewelke de CE1 ontstaan is uit de conditionnel temporel subjectif, die geleidelijk is geëvolueerd volgens het schema van de “bridging contexts” van Heine (2002).

Exploring and Understanding Neural Models for Clinical Tasks - Sushil Madhumita (12/03/2021)

Sushil Madhumita

  • 12/03/2021
  • Promotoren: Walter Daelemans (UAntwerpen) en Simon Šuster (University of Melbourne)

Abstract

We explore the use of deep learning techniques to understand medical text in tasksrelated to understanding patient conditions. Additionally, given the sensitive nature ofthe medical domain, an important requirement is to ensure that the models we trainare not biased due to nuances of the data and training algorithms. Hence, we alsodevelop several interpretability techniques to investigate the information learned bythese models during the training process.

In Chapter 1, we describe our efforts towards learning a holistic patient view in the formof semantic representations of patients from medical notes. We evaluate theserepresentations on the tasks of primary diagnostic and procedural category prediction,as well as in-hospital, 30 days post-discharge, and 1 year post-discharge mortalityprediction. We find that induced neural representations significantly outperformbag-of-features based patient representations when there are few cases to learn from.Furthermore, we find that the most frequent terms can be recovered most easily,although several medical terms are deemed highly important for classification.

In Chapter 2, we further develop techniques to find more complex feature-interactionpatterns learned by feedforward neural networks. We explore the use of if-then-elserules as explanation patterns, where these patterns are learned by training a rulelearning algorithm to estimate the outputs of a given neural network from a set of mostimportant features for it. With our method, we find that explanations learned in theform of hierarchical if-then-else rules have a high precision, although the recall can belower, providing an overall explanation faithfulness of nearly 80%. However, due to thehierarchical nature of the if-then-else rules, the explanations can sometimes becomecomplex to parse.

In Chapter 3, we adapt our methods to find non-hierarchical decision lists that cancapture longer contexts in the form of skipgrams and explain more complex sequentialarchitectures that use word embeddings as inputs. We successfully validate theexplanation patterns by evaluation on a synthetic textual dataset, which we mimic to besimilar to real clinical corpora. We find that the explanations correspond to the rules weuse to label the synthetic dataset. Later, we discuss the patterns learned by our modelson the real clinical task of sepsis classification at the time of patient discharge. From theexplanation rules, we find that lexical mentions of sepsis are often used for predictingsepsis when discharge notes are included, whereas more complex patterns involvingdifferent patient conditions are assessed when we exclude discharge notes fromanalysis. In a sentiment analysis task, we find that our methods are more scalable, andthat our explanations provide more context and are more faithful than existingstate-of-the-art rule-based explanation methods, although they also have higherexplanation complexity.

Finally, in Chapters 4 and 5, we analyze how existing state-of-the-art models trained ontask-specific data fail on several instances that require medical domain knowledge inthe tasks of medical language inference and relation classification between medicalentities mentioned in text. We focus on improving these models by supplementingdomain information from textual medical corpora. For the task of medical languageinference, we explore methods for augmenting in-domain knowledge both implicitly viafurther language modeling, and explicitly by adding relevant background information tothe instance. For relation extraction, we augment a feature that quantifies the pointwisemutual information (PMI) between medical entities to provide additional background forrelation classification. In both the setups, we find that this task of integratingbackground knowledge from textual corpora is extremely complex. We do not see anysignificant improvements with our methods, leaving this question open for furtherresearch.

To sum up, we develop several state-of-the-art techniques for medical textunderstanding, while also enabling understanding of these techniques to make theblack box algorithms more transparent, and to ensure capabilities for error analysis andfairness of these algorithms. This line of work will not only improve existing models fornatural language understanding, but will also promote a wider adoption of successfulmethods developed in academic research environments for more critical clinicaldecision making in hospitals.

Pragma-linguistic and socio-pragmatic perspectives on intercultural communicative competence within mobile and non-mobile higher education contexts: examining Turkish and Belgian higher education students’ speech act products in English - Diler Aba (29/01/2021)

Diler Aba

  • 29/01/2021
  • Promotor: Tom. F.H. Smits

Abstract

This study examines written speech act products of different student groups. First, mobile Turkish Erasmus students’ English requests and refusals are analyzed from the perspectives of modification, (in)directness and politeness. The analysis of the request data is made via the taxonomy of request modification in Halupka-Resetar (2014) while the analysis of the refusal data is made with the classification of refusals by Beebe et al. (1990). Certain changes and adaptations were applied to these taxonomies so that they served the goals of the present study and fit the characteristic features of all participant groups.

For the analysis of requests, the frequency and variety of various modification tools are examined. Modification is useful to increase politeness and to compensate for the face-threatening nature of speech acts. This analysis required the coding and enumeration of both internal and external modifiers in the students’ emails, text messages and/or social media messages. The interactions in these messages or emails take place between either unequal level interlocutors with high social distance or equal level interlocutors with low social distance. Some interactions require formal written communication (e.g. emails written to higher status interlocutors) with a high degree of imposition (i.e. request for action) while some others included informal written communication (e.g. text messages between friends) with a low degree of imposition (i.e. request for information). For the analysis of refusals, various direct and indirect refusal strategies and adjuncts are examined in these students’ written refusal products. Contextual variables are essential for appropriateness and politeness of the students’ refusals as well.

The quantitative analysis of the request and refusal data through the abovementioned taxonomies is supported with qualitative methods by evaluating the modification and (in)directness results according to the contextual variables specific in each writing task. This analysis is made to examine the politeness and appropriateness of the students’ speech acts. By analyzing several example texts from the students written products, the study increases understanding about mobile Turkish higher education students’ speech act performances in English. Results of this study are reported in case study one of chapter six.

A second case study in chapter six examines non-mobile Belgian higher education students’ written request and refusal products in English. The analysis methodologies in case study two are similar to case study one. However, the second case study not only carries out a performance analysis of the participants’ (Belgian higher education students) speech act products, but also compares these students’ speech acts with the Turkish students’ from case study one. This comparative analysis is made to add to the state of knowledge on non-native students’ requests and refusals in English. Previous research identified similarities between non-native speakers of English while generating various speech acts. In order to learn more about different non-native student groups’ speech act performances, case study two compares the Belgian and Turkish higher education students’ request and refusal products in English.

The final case study in this doctoral thesis investigates the relationship between academic mobility and speech act production. In order to do this, the study compares the request and refusal performances of mobile and non-mobile Turkish higher education students. Analysis methods in case study three are similar to the ones in the previous case studies in this thesis. In other words, modification use, (in)directness and politeness are examined in written speech act products of the participants. Case studies two and three carry out t-tests to compare the results between the groups. Additionally, correlation analyses are carried out to examine the relationship between academic mobility and speech act production.

The data in this research were elicited via discourse completion tasks developed in the present study, and adapted to both mobile and non-mobile higher education contexts. Chapter five explains the rationale in using such tasks in the present study. Furthermore, this chapter discusses the potential and limitations of employing discourse completion tasks in speech act research. The goal in examining the various participant groups’ speech act products in this study is to increase understanding about these students’ competence for interacting with people from other cultural and linguistic backgrounds.

Speech acts are essential linguistic components of pragmatic competence, which is fundamental to the success of intercultural interactions. Prior to presenting the case studies in chapter six, this thesis explains and discusses several variables and components that are essential for intercultural communicative competence. For example, chapter two discusses the relationship between pragma-linguistic competence and intercultural communication. Furthermore, chapter three examines exhaustively the relationship between intercultural communicative competence and academic mobility. Chapter four discusses the changes and developments in the higher education environments in Europe and focuses on the case of Turkey.

Finally, chapter four also presents a new intercultural readiness tool and discusses the best possible ways to examine students’ intercultural competencies in higher education. The studies in these chapters are published as research articles in various international journals, and their findings help to increase understanding about intercultural communicative competence, its various components and its relationship with students’ pragma-linguistic and socio-cultural skills from a wider perspective.

On periphrastic do and the modal auxiliaries. A connectionist approach to language change - Sara Budts (20/01/2021)

Sara Budts

  • 20/01/2021
  • Promotor: Peter Petré

Abstract

When no auxiliary is present, Present Day English requires a form of do to be used in negated and inverted clauses. Despite its pervasiveness today, auxiliary do (or periphrastic do) is only a sixteenth century innovation. Previous studies have related the rise of do to analogical pressure from the modal auxiliaries can, may, must, shall and will. The modals had grammaticalized in the twelfth century, but continued to develop until about 1550. To investigate the changing relation between do and the modals, I retrieved all instances of these forms from an 800 million word corpus that covers the late sixteenth and entire seventeenth century. Making use of Convolutional Neural Networks, I charted the prototypical usage contexts of do, as well as the contexts where do was functionally equivalent to one of the modals.

The results indicate that the modals are likely to have had an impact on the early rise of periphrastic do in the sixteenth century, but fail to account for the construction’s eventual regulation. By the end of the sixteenth century, do had come to serve as a marker of habituality and universality that typically occurred in generic statements to emphasise truthfulness. Late sixteenth century do essentially allowed the speaker to emphatically impose likelihood onto a proposition. This function may be considered a subtype of epistemic modality, which makes affirmative do functionally equivalent to the modals in certain contexts.

The seventeenth century, by contrast, is characterised above all by the divergence of do and the modals. While do was well on its way to join the paradigm of the modal auxiliaries, the verb’s modal semantics steadily decline from 1600 onwards. This developments is unlikely to be an effect of analogy with the modals.

From a theoretical angle, the present thesis aims to illustrate how a connectionist perspective and methodology can contribute to contemporary research in usage-based linguistics. A highly interdisciplinary framework of cognitive science, connectionism models cognition as a domain-general, inherently dynamic network of associations that accounts in a uniform way for language acquisition, processing and change. By adopting a connectionist methodology to study do in relation to the modal auxiliaries, I not only show how connectionist networks are data-analysis tools whose unprecedented flexibility enables us to make nearly holistic analyses in an automated fashion, I also argue how a connectionist stance allows us to include considerations from language processing and acquisition into research on diachrony.

Verandering in professioneel taalgebruik in stagelessen als mogelijke indicator voor de groei van de beroepsvaardigheid bij tweedegraads leraren in opleiding - Gosse Romkes (26/11/2020)

Gosse Romkes

  • 26/11/2020
  • Promotoren: Elke Struyf, Jef Verschueren en Sigurd D'hondt

Abstract 

In dit onderzoek staat het leerproces ‘Leren Lesgeven’ centraal zoals zich dat voltrekt bij studenten tijdens de initiële opleiding tot leraar.

Door inzicht te ontwikkelen in het leerproces ‘Leren Lesgeven’ van studenten kan het curriculum van de lerarenopleiding beter op studenten worden afgestemd. In het onderzoek is daarom op verschillende momenten de beroepsvaardigheid van leraren in opleiding in kaart gebracht. Een groep van bijna 30 voltijdse studenten uit de opleidingen Geschiedenis, Nederlands en Wiskunde hebben geparticipeerd in het onderzoek. Gedurende hun opleiding zijn lestranscripten gemaakt van de klasinteracties met leerlingen in stagelessen. In het onderzoek wordt dit professioneel taalgebruik van leraren in opleiding gezien als de uitdrukking van hun beroepsvaardigheid en markeert het professioneel taalgebruik het leerproces ‘Leren Lesgeven’ dat zij doorlopen.

De analyses van het professioneel taalgebruik is toegespitst op een twaalftal beroepsspecifieke lesgebeurtenissen of klasdiscoursevents die onderdeel uitmaken van een stageles. Deze klasdiscoursevents dragen bij aan de realisatie van taakgerichte-, lesorganisatorische en sociaal-emotionele functies van het klasdiscours in drie domeinen. Voor het leerdomein zijn de volgende klasdiscoursevents in de analyses betrokken: Huiswerkopgave bespreking, Vraagstellen, Instructie leerstof (uitleg), Leerstof recapitulatie, Scaffold lesopdracht, Lesopdracht bespreking uit het leerdomein van de klas. Bij het lesorganisatorisch domein zijn dat: Explicitatie van klas- en lesinformatie, Lesopdracht instructie, Regulatie van de eventuitvoering en Lesovergang maken. En bij het leefdomein zijn dat: Conversatie Leefdomein en Correctie deviant gedrag. De klasdiscoursevents zijn onderzocht op drie elementen: 1. De interactionele structuur, 2. De performatieve structuur, 3. De informatiestructuur.

In deelonderzoek I is het contrast in professioneel taalgebruik en beroepsvaardigheid tussen eerste en vierdejaars zichtbaar gemaakt. De nadruk lag op verschillen en overeenkomsten in beroepsvaardigheid als eerste glimp van het leerproces ‘Leren Lesgeven’. Het tweede deelonderzoek richtte zich op het volgen van elf studenten gedurende de vier stageperiodes van de lerarenopleiding. De longitudinale focus hier lag op eventuele veranderingen in het professioneel taalgebruik en de beroepsvaardigheid op bepaalde momenten in de opleiding om zich te krijgen op de inhoud en het verloop van het leerproces ‘Leren Lesgeven’. De uitkomsten uit beide deelonderzoeken laten een vergelijkbaar beeld zien. De overeenkomsten in de beroepsvaardigheid zijn groter dan de verschillen tussen eerste- en vierdejaars en de beroepsvaardigheid verandert in geringe mate gedurende de opleiding. De conclusie is dan ook dat er niet automatisch van uit kan worden gegaan dat het didactisch-lesorganisatorisch-pedagogisch repertoire verandert als gevolg van meer onderwijs/stage-ervaring en meer opleidingsjaren. In het onderzoek wordt uitgebreid stilgestaan bij mogelijke achtergronden van deze onderzoeksuitkomsten.

Dynamic Reading: the processing of orthographic features by dynamical systems - Stéphan Tulkens (11/09/2020)

Stéphan Tulkens

  • 11 september 2020
  • Promotoren: Walter Daelemans en Dominiek Sandra

Abstract

During single word reading, words are primarily identified by their visual properties: the letters or characters that make up that word. From a common sense point of view, this process of identification can be considered to be like looking up words in a dictionary. Several phenomena, however, point towards a much larger role for interactive, dynamic processing during visual processing: the identification of a word depends on the other words a person knows, and is not just the result of a visual stimulus. One of these phenomena is the neighborhood effect, which is the effect that words that look more like other words are read more quickly. For example, the word “BOOK” looks like several other words, such as “NOOK”, “ROOK”, “TOOK” and “LOOK”, while “SONIC” only looks like “TUNIC” and “TONIC”. All things being equal, “BOOK” is therefore read faster. The role of the neighborhood effect, and modeling the neighborhood effect using dynamic models of word recognition, is the topic of this thesis.

The thesis is divided into two parts, and is prefaced by an introductory theoretical chapter. The theoretical chapter mainly deals with the philosophical preliminaries of the work, and specifically with the notion of representation employed in many theories and models of word reading, and how a specific notion of representation is employed in many theories of word reading, including those governing the neighborhood effect.

The first part of this thesis deals with measuring the neighborhood effect, and specifically with the notion of similarity employed in the definition of the neighborhood effect. In short, there exist many featurizations of orthography, also called orthographic codes, all of which result in differences in how words can be said to be orthographically similar. According to some feature sets, the words “TREE” and “THREE”, for example, are very similar, while to others, these words are very dissimilar. In the first chapter, we discuss wordkit, a Python package containing implementations of all these feature sets, along with tools for accessing popular corpora. This toolkit thus facilitates research in word reading.

In the second chapter, we compare several feature sets, and introduce a new metric for comparing them, which we call Representation Distance (RD). This metric is the sum of the distances to the 20 closest neighbors for an arbitrary vector space. Previous metrics only operated on string representations, and therefore could not be used to compare theories of orthographic representation. In a regression analysis using French, Dutch and British English corpora, we show that feature sets that are more flexible, i.e., feature sets that assign higher similarity to THREE and TREE, explain less variance in lexical decision latencies. We also show that RD, the newly proposed metric, can be used to analyze hidden state matrices of Multi-Layer Perceptrons.

In the third chapter, we expand this analysis, and optimize the different feature sets by jointly optimizing the number of nearest neighbors taken into account in the calculation of RD, and the parameters of the feature sets. We apply this to five alphabetic languages: the three languages used in the previous study, as well as American English and Spanish. We show that all results of the optimizations are in agreement across these languages. We also introduce a feature weighting scheme based on a discrete version of negentropy, the complement of entropy. Finally, we also present results on Hangul, the alphabet of Korean. These results diverge from the results on the other alphabetic languages, and, as such, raise questions of whether the way we measure the neighborhood effect is specific to alphabetic languages.

The second part of the thesis deals with the dynamic models that process these orthographic features. In this part, we specifically study the Interactive Activation (IA) model, a dynamic model of word recognition first published in 1981. In a first chapter, we introduce Metameric, a toolkit for the IA model and related models, together with some empirical work demonstrating some of the assumptions inherent in the IA model and IA networks. We specifically show that the assumption that IA models can only use words of a specific length is false, and show how the model can be extended to use words of any length. In the second chapter, we present a working version of the BIA+ model, a bilingual version of the IA model. This model, although very popular, has only been available as a theoretical model, and has not been implemented before. Using a process called rational reconstruction, we use the theoretical version of the BIA+ model to create an implemented version. We show that one of the main theoretical commitments of BIA+, that language information is explicitly represented, is not backed up by the implemented version of the BIA+ model. In the final chapter, we introduce a new model, called SIMBL, which is a simple dynamic model of lexical decision. This model is more flexible than the IA model in terms of which orthographic features it accepts, but still can explain some common results in lexical decision experiments.

Finally, we present a short conclusion in which we discuss how these dynamic models could be extended to a fully visually grounded model of reading.

 

Tijdens het lezen worden woorden geïdentificeerd door middel van hun visuele eigenschappen, zoals de letters of andere karakters waaruit woorden bestaan. Dit proces van visuele identificatie zou men metaforisch als een soort van toegangscode kunnen zien: de reeks karakters die een woord voorstelt is een sleutel van een bepaalde kluis, die de betekenis en associaties van een woord bevat. Verschillende fenomenen spreken dit tegen, en laten zien dat dit proces van toegang eerder dynamisch en interactief is. De identificatie van een woord hangt niet puur af van de visuele stimulus, maar wordt beïnvloedt door andere woorden die de lezer kent. Een van deze fenomenen is het zogenaamde neighborhood effect: woorden die meer op andere woorden lijken worden sneller geïdentificeerd dan woorden. Dit fenomeen is waarschijnlijk het gevolg van competitie of co-activatie tussen woord-representaties. Het bestuderen van dynamische modellen die dit fenomeen kunnen verklaren, is het onderwerp van deze thesis.

De thesis is verdeeld in twee delen, en wordt voorafgegaan door een theoretisch kader. Het theoretisch kader bevat, naast een introductie, ook de filosofische veronderstellingen van het werk. Hierin besteden we met name aandacht aan de notie representatie, een term die in de beschrijving van veel modellen gebruikt wordt, maar zelden volledig uitgelegd wordt.

Het eerste deel van de thesis gaat over het neighborhood effect als zodanig, en specifiek over de notie van orthografische gelijkenis die in de definitie van het effect gebruikt wordt. In het kort zijn er veel verschillende mogelijkheden om orthografie te representeren. Elk van deze mogelijkheden noem je een feature set. Feature sets verschillen dus in woorden op elkaar lijken. Volgens sommige feature sets lijken de woorden DRIE en DIE bijvoorbeeld veel op elkaar, voor andere dan weer niet. In het eerste hoofdstuk van dit deel bespreken we wordkit, een softwarepakket dat veel van deze feature sets implementeert, en daarmee de reproduceerbaarheid van experimenten vergroot, omdat het softwarepakket hergebruikt kan worden voor nieuwe experimenten. In het tweede hoofdstuk vergelijken we een aantal van deze feature sets, en introduceren een nieuwe metriek om ze te vergelijken. Deze metriek noemen we Representation Distance (RD). De RD van een bepaald woord is gedefiniëerd als de afstand tot de twintig dichtstbijzijnde andere woorden in een arbitraire vectorruimte die gedefinieerd wordt door de feature set. Eerdere metrieken opereerden alleen op string representaties, en konden daarom niet gebruikt worden om arbitraire feature sets met elkaar te vergelijken. Met RD laten we zien dat feature sets die meer orthografische flexibiliteit toelaten tegelijkertijd minder variantie verklaren in de reactietijden van lexicale decisie experimenten in drie verschillende alfabetische talen. Daarnaast laten we zien dat RD gebruikt kan worden om inzicht te krijgen in hoe een neuraal netwerk, zoals een Multi-Layer Perceptron (MLP), de woorden verwerkt. In het derde hoofdstuk breiden we de analyse van het vorige hoofdstuk uit, en optimaliseren we de verschillende feature sets door tegelijk het aantal woorden dat in de berekening van RD gebruikt wordt, en de parameters van deze feature sets, te optimaliseren.

Daarnaast introduceren we twee nieuwe corpora, en laten we zien dat de resultaten die we vinden identiek zijn voor al deze talen, wat aantoont dat onze resultaten niet taalspecifiek zijn. Daarnaast introduceren we een techniek die individuele features van de feature set weegt. Deze is gebaseerd op een discrete variant van negentropie, het complement van entropie. Tenslotte presenteren we ook de eerste resultaten op Hangul, het Koreaanse alfabet. De resultaten voor het Koreaans divergeren van die van de andere vijf alfabetische talen. Dit laat zien dat het neighborhood effect wellicht geen algemeen cognitief effect is, maar eerder specifiek is aan alfabetische talen.

Het tweede deel van deze thesis gaat over dynamische modellen die de orthografische features uit het eerste deel verwerken. In dit deel bestuderen we specifiek het Interactive Activatie-model (IA), een dynamisch model van woordherkenning uit 1981. In een eerste hoofdstuk introduceren we Metameric, een simulator en tool waarmee IA en andere modellen makkelijk gereproduceerd kunnen worden. Daarnaast bevat het hoofdstuk enkele algemene empirische resultaten over het IA-model en IA-netwerken in het algemeen. In het tweede hoofdstuk van dit deel bespreken we het BIA+ model, een bilinguaal IA-model. Het BIA+ model had tot de publicatie van dit hoofdstuk nog geen implementatie, alleen maar een theoretische beschrijving. We beschrijven de implementatie van het BIA+ model, en de moeilijkheden die we in het creëren van deze implementatie tegenkwamen. Een vergelijking van het theoretische en geïmplementeerde model laat zien dat het BIA+ model een te grote nadruk legt op de rol van taligheid. Voor de simulatie van bilinguale effect is, althans in het framework van BIA+, geen notie van taal nodig. In het laatste hoofdstuk van dit deel, en tevens het laatste hoofdstuk van deze thesis, introduceren we SIMBL, een algemeen dynamisch model dat, in tegenstelling tot IA, niet beperkt is tot het gebruiken van een enkele soort orthografische representatie. We laten zien dat SIMBL verschillende effecten in lexicale decisie experimenten kan verklaren.

Identifiability and intelligibility of the speech of children with a cochlear implant: a comparison with normally hearing children and children with an acoustic hearing aid - Nathalie Boonen (08/07/2020)

Nathalie Boonen

  • 8 juli 2020
  • Promotoren: Steven Gillis en Hanne Kloots

Abstract

This dissertation investigated the speech and language of hearing-impaired primary school aged children. They were born with a hearing impairment and their hearing was aided very early on in life. Two issues pertinent to their long-term speech outcomes were studied. The first one related to the speech intelligibility of these hearing-impaired children: is their speech as intelligible as that of their normally hearing peers? The second one related to the children’s speech identifiability and addressed the question whether the speech of hearing-impaired (HI) children is distinguishable from the speech of normally hearing (NH) peers? For both topics, the speech of hearing-impaired children with cochlear implants (CI) and acoustic hearing aids (HA) was compared to that of NH peers.

The studies on speech intelligibility resulted in two important findings. On the one hand, it was found that the speech intelligibility of seven-year-old children with CI had not yet reached NH children’s level. As a group, their intelligibility scores were significantly lower than those of NH peers. On the other hand, intelligibility scores exhibited substantial individual variation: some children with CI scored well within the range of children with NH, while others clearly lagged behind.

In the studies on speech identifiability, it was found that listeners reliably differentiated the speech of NH and HI children, indicating that their speech was indeed identifiable. This result indicated that HI children’s speech contains marking cues that are salient to listeners. Future research has to investigate which cues are most salient to listeners. In the group of HI children, the results showed an effect of the length of device use for children with HA as well as for children with CI. Interestingly, this effect was far more pronounced in the latter group. Moreover, the results indicated that experience with the speech of (HI) children or familiarity with the native language of the child did not affect the judgements considerably.

The re-autonomization of the modal auxiliaries in Dutch - Wim Caers (23/06/2020)

Wim Caers

  • 23 juni 2020
  • Promotor: Jan Nuyts

Abstract

​Deze thesis beoogt een grondige analyse van het herautonomiseringsproces van de Nederlandse modale hulpwerkwoorden. Ze heeft tot doel de timing, de grammaticale en semantische eigenschappen, alsook de factoren die een rol spelen in het ontstaan en de ontwikkeling van de nieuwe autonome gebruiken van de modale werkwoorden te onderzoeken. De studie is corpusgebaseerd en maakt gebruik van representatieve steekproeven uit drie verschillende stadia van het Nieuw Nederlands, te weten: Vroegnieuwnederlands, Nieuwnederlands en huidig Nederlands. Ze toont aan hoe kunnen, mogen en moeten in het huidige Nederlands in toenemende mate zelfstandig gebruikt worden in de zin (d.w.z. zonder de aanwezigheid van een ander hoofdwerkwoord), terwijl in hoeven de toename van nieuwe autonome gebruiken al plaatsvond vanaf het Nieuwnederlands. Ze laat ook zien dat de nieuwe autonome gebruiken geen voortzetting zijn van de oorspronkelijke hoofdwerkwoorden, maar integendeel ontstaan zijn uit de zuiver hulpwerkwoordelijke gebruiken (via een tussenfase met een geïmpliceerd hoofdwerkwoord), hetgeen suggereert dat we hier te maken hebben met een geval van (collectieve) degrammaticalisatie. Vanuit grammaticaal perspectief toont de studie aan dat de ontwikkeling van de nieuwe autonome gebruiken langs twee alternatieve paden verloopt. Enerzijds zijn er gevallen waarvan het eerste argument verwijst naar een stand van zaken. Anderzijds zijn er gevallen met een (pro)nominaal eerste argument. Vanuit semantisch oogpunt wordt aangetoond dat de nieuwe autonome gebruiken (zeer) subjectieve en intersubjectieve betekenissen uitdrukken (vergelijkbaar met de hulpwerkwoorden), en geen lexicale betekenissen (in tegenstelling tot hun oorspronkelijke hoofdwerkwoorden). Wat de mogelijke verklaringen voor het herautonomiseringsproces betreft, toont de studie aan dat hoewel directionele elementen een rol spelen in het stimuleren van nieuwe autonome gebruiken van het type met een (pro)nominaal eerste argument (net zoals in andere Germaanse talen), zij niet de bepalende factor zijn die de algemene herautonomiseringstendens in het Nederlands veroorzaakt. Negatie daarentegen lijkt wel een rol te spelen in het algehele herautonomiseringsproces, maar hoe precies blijft vooralsnog onduidelijk.

'Digital Nature for healthcare': Onderzoek naar beeldende kunst en gezondheidszorg om stress bij kinderen te verminderen - Ludivine Lechat (19/06/2020)

Ludivine Lechat

  • 19 juni 2020
  • Promotoren: Koen Norga, Monica Dhar en Tom De Smedt

Abstract

Dit doctoraat beoogt het promoten van onderzoek naar beeldtaal voor positieve artistieke interventies in de zorgsector, om de levenskwaliteit van jonge patiënten te verbeteren.

In overleg met kinderen, verplegend personeel en opvoeders gaat Ludivine Lechat op zoek naar visuele voorkeuren en narratieve concepten die het mentale welzijn van kinderen zou kunnen bevorderen. Dit om inzichten te verwerven voor het ontwerpen van ontspannende illustraties voor nieuwe media applicaties.

Het gehele onderzoeksproces heeft bijgedragen aan de ontwikkeling van creatieve product prototypes dat gericht zijn op stressverlagende toepassingen binnen de zorgsector.

“Digital Nature for healthcare” is een poging om via beeldende kunsten op een zinvolle manier bij te dragen aan een sector die ons uiteindelijk allemaal aanbelangt: gezondheidssector.

Special passives across the lifespan: cognitive and social mechanisms - Lynn Anthonissen (08/06/2020)

Lynn Anthonissen

  • 8 juni 2020
  • Promotoren: Peter Petré, Tanja Mortelmans en Hans-Jörg Schmid

Abstract

Linguists have typically studied constructional change at the aggregate level of speech communities, yet key mechanisms of change (e.g. analogy, automation, schematization) operate within the minds of individual language users. While this is tacitly assumed in the usage-based literature, longitudinal studies investigating to what extent and under which conditions speakers adapt to ongoing change are scarce, especially in the domain of syntax. The present study aims to make headway in this area by addressing three fundamental issues relating to individuality in language change, i.e. how variation and change at the individual level interact with change at the community level (i), how much innovation and change is possible across the adult lifespan (ii) and to what extent related patterns are associated in individual cognition (iii).

To gain insight into the cognitive and social mechanisms that drive linguistic change, the present study has selected two interrelated syntactic constructions that undergo significant change in the history of English: the prepositional passive (e.g. he was laughed at) and the Nominativus cum Infinitivo or ‘NCI’ in short (e.g. he is said to be a thief). The expansion of these special passive constructions in Early Modern English presents itself as an interesting case study to explore individual trajectories in language change against the backdrop of temporal language dynamics in the community. The corpus that was established to this purpose contains lifespan data from 50 prolific seventeenth-century authors, who hail from various interconnected communities of practice.

As to the first issue specified above (i), the analysis shows that population-level trends often conceal a great deal of variation at the individual level, and may even be in contradiction with individual trajectories. While variation between individuals is in part constrained by social factors, the amount of individuals’ unique residue is not trivial. It is shown that such variation can play a meaningful role in advancing ongoing changes as small shifts in frequency and productivity can increase in magnitude over time, leading to what is eventually recognized as language change. The rate of change may also be accelerated if at least a significant minority of speakers undergoes lifespan change in the direction of the community trend.

The study also provides novel insights into the issue of lifespan change in the domain of syntax (ii). While most authors do not appear to be particularly sensitive to ongoing change, some authors do assimilate their usage across the lifespan. Concerning the usage rates of the special passives, fairly extreme lifespan increases are documented, which shows that cognitive routinization may significantly reinforce the use of existing patterns in adulthood and even old age. It is also demonstrated that when a syntactic construction (e.g. the NCI construction) has two different semantic-pragmatic functions, their relative weight may shift significantly across the lifespan. Attested lifespan changes are typically aligned with the community trend, yet speakers may also behave in unpredictably ways, for instance by crystallizing lexically filled patterns at lower levels of the constructional hierarchy which are not the dominant patterns in the community at large.

Finally, this study investigates to what extent related patterns such as the special passives are associated in individual cognition (iii). Modeling techniques suggest strong degrees of grammar-internal coherence, with special passives being persistently associated in individual usage. At the aggregate level, however, the association appears to become weaker and in a couple of authors the special passives are found to diverge with age. It is argued that these patterns of dissociation reflect long-term diachronic changes in the usage profiles of the two special passives: the use of the NCI becomes saturated and specializes toward the expression of evidentiality, whereas the prepositional passive continues to recruit new and more complex items. Functionally, these changes are motivated by an inherent disparity in semantic scope; cognitively, they reflect the differential effects of what becomes routinized and entrenched.

Metacognitive awareness and academic literacy skills in English as a foreign language: The case of summarising - Marina Vulovic (20/12/2019)

Marina Vulovic

  • 20 december 2019
  • Promotor: Kris Van de Poel

Abstract

A range of studies has demonstrated that the transition from secondary to higher education (HE) remains a challenging experience for first-year students. A primary factor that has an impact on students’ acculturation to the new learning and teaching environment, and by default on their academic success, is their mastery of general and discipline-specific academic discourse. In other words, novice university students are required to learn to effectively and efficiently communicate in a manner established by their academic community by engaging with academic discourse in all its diversity. Not surprisingly, many students experience the process of academic acculturation as challenging (at best), particularly due to its implicit character and unwritten rules, which as a whole seems to create tensions and discrepancies in students’ understanding of the academic expectations of their new community. An effective command of disciplinary discourse is, nonetheless, critical for the students’ progression in HE and can be attained through their academic literacy development whether implicit or explicit. 

In this study, effective and integrated reading and writing are understood as central academic literacy practices through which students learn and develop their knowledge within disciplines and master disciplinary discourse. As written assignments remain the main form of assessment in HE, writing is regarded as a dominant literacy practice. However, at the very least, academic reading is an equally important literacy practice; and yet, it is often taken for granted and in turn given less attention in support materials for students' academic literacy development. Thus, a key component in the teaching and learning of English for Academic Purposes (EAP) —learning to write from reading input— academic reading remains a prevailing issue. More specifically, core academic literacy practices such as summarising and paraphrasing, in particular in a foreign and second language, continue to be challenging for first-year students and often for senior undergraduates as well. Consequently, it is important to rethink teaching and learning support 

for first-year students’ academic literacy skills development in a responsible and informed way. 

As research up to date has indicated a positive correlation between metacognition and academic performance, this study takes the perspective that first-year students can be supported in their academic literacy skills development by increasing their metacognitive awareness of their competencies. In other words, raising students’ awareness of their strengths and weaknesses can help them better assess their own mastery of academic tasks, focus their skills development and perfect their knowledge. 

With the aim to further inform the design of first-year students’ academic literacy support, this thesis examines students’ strengths and weaknesses within core academic literacy practices at the beginning of HE, as well as their understanding of those competencies. The research took place at the University of Antwerp in Belgium among first-year English majors for whom English is a foreign language. In the exploratory study, students (N = 116) discussed in small groups how they had experienced academic reading (in relation to writing) during their first three months at university. In the main study, summarising was focused on. This is a skill that students make use of when checking their reading comprehension and analysing a text, and, at the same time, it represents a prototypical academic literacy task. In order to enrich the preliminary insights, a new group of students’ (N = 109) perceived and actual summarising skills (and the relationship between them) were studied by making use of a mixed-method research design. 

The findings of this study confirm that first-year students’ weaknesses in core academic literacy skills, such as summarising, are persistent. More in particular, filtering primary from secondary information and paraphrasing remain challenging, which indicates that students need additional support in both their receptive and productive skills in summarising. Moreover, the students’ misunderstanding or lack of understanding of their competencies can have a detrimental impact on their academic literacy development. More specifically, discrepancies have been identified in students’ awareness of their summarising skills in English as a foreign language (EFL): students tend to overestimate or underestimate their summarising skills. The aforementioned discrepancies may explain first-year students’ struggle in mastering academic discourse. Consequently, this research suggests making 

use of the identified discrepancies to help raise students’ awareness of their competencies in the early stages of higher education in order to additionally support their skills development. Moreover, the data show that the approach underlying the task used for the purposes of this study is useful in order to gain an insight into students’ understanding of reading and writing materials and can in turn inform syllabus design. Since a one-size-fits-all approach has been contested, suggestions are made as to how both approach and findings are relevant for local as well as wider contexts.

Learner Autonomy in the Acquisition of Advanced Academic Literacy - Elke Ruelens (20/12/2019)

Elke Ruelens

  • 20 december 2019
  • Promotor: Kris Van de Poel

Abstract

The role of learner autonomy does not feature explicitly in approaches to advanced academic literacy instruction. Throughout their educational programme, however, learners are required to autonomously advance their literacy. Both a thorough understanding of what academic literacy entails and the ability and willingness to manage their learning, i.e., from analysing their learning needs to monitoring the learning process and critically evaluating its outcome, are prerequisites for fulfilling this expectation. 

Given the weight that has been attributed to learner autonomy, it has been generally recognised that learners need support to (learn to) take responsibility for their learning. Not only is it essential that learners learn how to manage their learning process (i.e., they have to develop self-regulative competence), they should also feel competent and confident that they can successfully do so (i.e., the motivational aspect of learner autonomy). 

One powerful approach to support the development of learner autonomy is strategy-based instruction, in which learning strategies and content teaching are integrated. Strategy training has been shown to positively affect learners’ self-regulative competence as well as their writing ability. It seems, however, that the effect of strategy training on learners’ self-efficacy beliefs has not been examined thoroughly, which is why this study explored how university language students’ self-efficacy beliefs evolve throughout an academic literacy course. 

To this end, a strategy-based training package, systematically introducing metacognitive and social strategies, was designed. The trained strategies were: analysing learning needs and setting goals, planning the learning process, seeking social support, monitoring the learning process and evaluating learning process and outcome. The package was embedded in a second-year academic literacy course of which the objectives were (1) to improve the participants’ discourse knowledge and ability to write an academic text, and (2) to enhance their self-efficacy beliefs regarding the trained learning strategies. In two consecutive years, the effectiveness of the training package, in terms of raising students’ self-efficacy beliefs, was examined (2016-2017: n = 55 and 2017-2018: n = 62).

At the beginning and end of each course, data were gathered about the participants’ self-beliefs as writers of academic texts and about their self-efficacy beliefs concerning metacognitive and social learning strategies. Comparative analyses revealed that, overall, the training fulfilled its two objectives. The course had a significant positive effect (1) on the respondents’ self-beliefs about themselves as writers, (2) on their ability to write an academic text and (3) on their self-efficacy beliefs about learning strategies relating to needs analysis, help-seeking and evaluation. The analyses also uncovered some challenges for strategy-based academic literacy instruction.  First, strategy training should include learning activities that give students impetus to consciously pay attention to the learning process and not only to the outcome. In other words, students should get the opportunity to plan, manage and monitor their learning process for them to feel more efficacious. Second, the results testified to a need for additional awareness raising concerning time management, since many respondents attributed shortcomings in their planning to external factors, thus failing to recognise their own share in time management.

This study has attested to the potential of strategy-based instruction to improve language students’ self-efficacy beliefs concerning the use of learning strategies. When the findings are considered in light of earlier research, it can be concluded that strategy-based instruction is an effective autonomy-supportive approach provided that its design caters to the participants’ preparedness for self-regulation. 

Measurements of Grammaticalization: Developing a quantitative index for the study of grammatical change - David Correia Saavedra (07/06/2019)

David Correia Saavedra

  • 7 juni 2019
  • Promotoren: Martin Hilpert en Peter Petré
  • Dubbeldoctoraat UAntwerpen - Université de Neuchâtel (Zwitserland)

Abstract

There is a broad consensus that grammaticalization is a process that is gradual and largely unidirectional: lexical elements acquire grammatical functions, and grammatical elements can undergo further grammaticalization (Hopper and Traugott 2003). This consensus is based on substantial cross-linguistic evidence. While most existing studies present qualitative evidence, this dissertation discusses a quantitative approach that tries to measure degrees of grammaticalization using corpus-based variables, thereby complementing existing qualitative work. The proposed measurement is calculated on the basis of several parameters that are known to play a role in grammaticalization (Lehmann 2002, Hopper 1991), such as token frequency, phonological length, collocational diversity, colligate diversity, and dispersion. These variables are used in a binary logistic regression model which can assign a score to a given linguistic item that reflects its degree of grammaticalization.

Grammaticalization can be conceived in synchrony and in diachrony. The synchronic view of grammaticalization is concerned with the fact that some items are more grammaticalized than others, which is commonly referred to as gradience. The diachronic view is concerned with the development of grammatical elements over time, whereby elements become increasingly more grammatical through small incremental steps, which is also known as gradualness. This dissertation proposes studies that deal with each of these views.

In order to quantify the gradience of grammaticalization, data from the British National Corpus is used. 264 lexical and 264 grammatical elements are selected in order to train a binary logistic regression model. This model can rank these items and determine which ones are more lexical or more grammatical. The results indicate that the model makes successful predictions overall. In addition, generalizations regarding grammaticalization can be supported, such as the relevance of key variables (e.g. token frequency, diversity to the left of a given item) or the ranking of morphosyntactic categories as a whole (e.g. adverbs are on average in between the lexical and grammatical categories).

The gradualness of grammaticalization is investigated using a selection of twenty elements that have grammatical and lexical counterparts in English (e.g. keep). The Corpus of Historical American English (1810s-2000s) is used to retrieve the relevant data. The aim is to check how the different variables and the grammaticalization scores develop over time. The main theoretical value of this approach is that it can offer an empirically operationalized way of measuring unidirectionality in grammaticalization, as opposed to a more qualitative observation based on individual case studies.

Sociale patronen in taalgebruik op sociale media: De invloed van leeftijd, gender en sociale-klasse-indicatoren op de informele online schrijfpraktijk van tieners - Lisa Hilte (29/05/2019)

Lisa Hilte

  • 29 mei 2019
  • Promotoren: Reinhild Vandekerckhove en Walter Daelemans

Abstract

Dit proefschrift behandelt sociale variatie in de informele online schrijftaal van tieners. We bestuderen 400000 chatberichten geproduceerd op de populaire sociale media Facebook Messenger en WhatsApp door meer dan 1000 Vlaamse tieners. De talige variabelen zijn verschillende ‘afwijkingen’ van de formele Nederlandse schrijftaal (bv. typische chattaalkenmerken zoals emoji, of traditionele niet-standaardtaligheid zoals regionaal taalgebruik), maar ook algemenere teksteigenschappen (bv. gemiddelde zinslengte). We onderzoeken de talige impact van leeftijd, gender en sociale-klasse-indicatoren als de studierichting van de jongeren en het beroep van hun ouders.

Significante stilistische variatie wordt met generalized linear mixed models waargenomen voor alle sociale variabelen, die bovendien ook met elkaar interageren. Oudere tieners, theoretisch opgeleide ASO-leerlingen en jongeren uit de bovenklasse produceren standaardtaligere chatberichten dan respectievelijk jonge tieners, praktisch opgeleide BSO-leerlingen en jongeren uit de arbeidersklasse. Leerlingen die een ‘hybride’ TSO-opleiding volgen, nemen geen talige tussenpositie in, maar hanteren een meer variabele online schrijfstijl. Jongens en meisjes blijken verschillende soorten niet-standaardtaligheid te verkiezen: jongens prefereren traditionele niet-standaardtalige kenmerken (bv. regionaal taalgebruik) terwijl meisjes meer de expressieve-typografische markers hanteren die eigen zijn aan het online genre (bv. emoji).

Belangrijk is dat jongeren uit het BSO en/of uit arbeidersmilieus minstens evenzeer aansluiting vinden bij de online schrijfcultuur als hun leeftijdsgenoten met een andere achtergrond: zij benutten volop de nieuwe communicatieve mogelijkheden van digitale media.

Vervolgens trachten we met classificatie-algoritmes de studierichting van middelbare scholieren te voorspellen a.d.h.v. hun socialemediaberichten. Het onderscheid tussen ASO- en BSO-leerlingen blijkt relatief makkelijk te maken o.b.v. specifieke voorkomens van stilistische chattaalkenmerken, maar het herkennen van TSO-studenten is moeilijker.

Tot slot peilen we via een enquête naar de attitudes van Vlaamse tieners m.b.t. het online taalgebruik van hun leeftijdsgenoten. De resultaten leren dat tieners zich goed bewust zijn van gender- en leeftijdsgerelateerde sociolinguïstische variatie. Wat studierichting betreft, blijkt dit bewustzijn echter nagenoeg afwezig. Een sterke registergevoeligheid wordt – ongeacht het profiel van de jongeren – waargenomen m.b.t. prototypische kenmerken van online schrijftaal, alsook een sterk gedeeld aanvoelen van potentiële negatieve connotaties van bepaalde chattaalkenmerken.

Dit proefschrift biedt een genuanceerd inzicht in sociale variatie in online jongerentaal, in het bijzonder d.m.v. de toevoeging van sociale-klasse-indicatoren en interacties tussen sociale variabelen. De waargenomen registergevoeligheid van tieners biedt een positieve bijdrage tot het debat over de mogelijke negatieve invloed van informeel online taalgebruik op traditionele geletterdheid.

Des deutschen Dichters Sendung. Die Sakralisierung von „Führer", „Reich" und „Volk" in der national–sozialistischen Dichtung: Heinrich Anacker, Gerhard Schumann und Herybert Menzel - Anneleen Van Hertbruggen (20/05/2019)

Anneleen Van Hertbruggen

  • 20 mei 2019
  • Promotor: Arvi Sepp

Abstract

Een opvallend religieus discours op alle niveaus van het nationaalsocialistische systeem – bijvoorbeeld in de politieke cultus, de propaganda en toespraken van Hitler en diverse andere hooggeplaatste nazifunctionarissen – gaf reeds in het „Derde Rijk“ aanleiding tot het beschrijven van het nationaalsocialisme als een zogenaamde „politieke religie“. Deze studie onderzoekt de tekstuele representatie van deze „politieke religie“ in affirmatieve nazipoëzie en in het bijzonder de sacralisering van drie centrale ideologemen: „Führer“, „Reich“ en „Volk“. Met dit doel belicht het theoretische luik van deze studie de relatie tussen het nationaalsocialisme en religie enerzijds en nationaalsocialisme en poëzie anderzijds. De literaire analyse focust op de kristallisatie van het religieus discours in de affirmatieve nazipoëzie om zo te onderzoeken hoe politieke en profane kernideeën zoals „Führer“, „Reich“ en „Volk“ naar een sacraal niveau getild worden.

Vanuit een theoretisch perspectief wordt eerst de vraag onderzocht in hoeverre het begrip „politieke religie“ daadwerkelijk toepasbaar is op het nationaalsocialistische regime. Als inleiding werpt deze studie eerst een blik op het totalitarismeonderzoek sinds de Tweede Wereldoorlog, om nadien dieper in te gaan op het traditionele onderzoek naar „politieke religies“. Hoewel de basis voor dit onderzoek al in de jaren dertig van de vorige eeuw werd gelegd, ontstond er pas vanaf de jaren negentig een breed en genuanceerd onderzoeksveld. Dit proefschrift beschrijft niet enkel de problematiek wat betreft het definiëren van de „politieke religie“, maar pleit bovendien voor een verruiming van het concept „religie“ zelf. Op deze manier biedt dit proefschrift een nieuwe visie op „(politieke) religie“, die zo als hanteerbaar concept gebruikt kan worden voor de beschrijving van totalitaire regimes zoals het nationaalsocialisme. Nog op theoretisch niveau probeert deze studie een antwoord te bieden op de vraag: „Wat is nationaalsocialistische poëzie?“ Om tot een genuanceerde en ideologiekritische interpretatie van deze politieke poëzie te komen, die net wegens haar utilitaire functie door de literatuurwetenschap na de Tweede Wereldoorlog grotendeels genegeerd werd, beroept deze studie zich op de tekstsociologische benadering van de Oostenrijkse literatuurwetenschapper Peter Zima. 

De concrete tekstanalyse en dus de interpretatie van het religieuze discours in affirmatieve nazipoëzie vormt de eigenlijke kern van dit onderzoek. Hoewel dit onderzoek aantoont dat de ideologemen „Führer“, „Reich“ en „Volk“ reeds over een zeker inherent sacraal potentieel beschikken, worden ze in de affirmatieve nazipoëzie doelgericht in verbinding met een religieus discours verder religieus opgewaardeerd. Doordat de analyse telkens op de sacralisering van één van de drie ideologemen in één specifieke gedichtbundel focust, valt het tekstanalytische luik van dit proefschrift uiteen in drie thematische zwaartepunten. Het eerste deel onderzoekt de voorstelling van Adolf Hitler – de „Führer“ – als een soort „Messias“ in de gedichtbundel Die Fanfare. Gedichte der deutschen Erhebung (eerste uitgave in 1933) van Heinrich Anacker (1901-1971). De interpretatie gebeurt tegen de achtergrond van het zogenaamde „politiek messianisme“ en de typerende messiaanse verlossingsbehoeften in het door crisis geteisterde Duitsland na het einde van de Eerste Wereldoorlog. Voor de sacrale vormgeving van „Reich“ in de anthologie Die Lieder vom Reich (1935) van Gerhard Schumann (1911-1995) keert dit proefschrift terug naar de theologische wortels van „het Derde Rijk“, met name naar de verdeling van de geschiedenis in drie tijdperken door de middeleeuwse monnik Joachim van Fiore (ca. 1130-1202). Hoewel „Volk“ het meest profane concept van de drie lijkt te zijn, wordt ook „Volk“ in de gedichtbundel Gedichte der Kameradschaft (1936) van Herybert Menzel (1906-1945) omgeven door een religieus aura. In het bijzonder wanneer individuen als martelaren worden voorgesteld en – in mindere mate – wanneer de maatschappelijke rol van vrouwen en mannen religieus wordt geduid, verheft ook dit „Volk“ zich tot een sacraal opgewaardeerd concept. Net door deze doelgerichte sacralisering van „Führer“, „Reich“ en „Volk“ beschouwt dit proefschrift deze drie ideologemen als het ware als „geloofsartikelen“ van de nationaalsocialistische (politieke) religie.

The Typology of Modality in Modern West Iranian Languages - Sepideh Koohkan (17/04/2019)

Sepideh Koohkan

  • 17 april 2019
  • Promotoren: Jan Nuyts en Arsalan Golfam

Abstract

Modality concerns with the modifications and semantic changes, which the speaker  makes in the proposition to indicate his/her commitment and assessment to the state of  affairs. On the other hand, typology deals with the varieties in languages to achieve  generalizations cross-linguistically. This thesis studies modality, as a semantic notion and  typology, as mostly a formal category, in eleven New West Iranian languages, including  Balochi (Bamposht), Gerashi, Gilaki (Shaft), Hawrami (Hawraman Takht), Kahangi,  Kordish (Sanandaj), Lori (Balaguariveh), Persian, Semnani, Tati (Takestan) and Vafsi  based on Nuyts (2005 and furthermore). The main goal of this dissertation is to examine  the possibility of categorizing languages based on a semantic feature, which is modality  here, and to discover the differences between this type of classification of languages and  the other available categorizations which mostly have a morphological or syntactic basis.  

The results reveal that all these languages, enjoy different types of expressions to express  modality, including modal auxiliaries, nouns, adjectives, adverbs and main verbs (mostly  complex verbs). However, this enjoyment is systematic, i.e. first, in all above languages,  these are the modal auxiliaries which not only have a high frequency in the languages,  but they are also the more native elements. Secondly, some languages, including Balochi  (Bamposht dialect) and Hawrami (Hawraman Takht dialect) apply adverbs to express  those dimensions of modality which prototypically is on modal auxiliaries in other  languages to express and that is why they contain less modal auxiliaries comparing to  other languages. Finally, modal nouns and adjectives (and consequently modal main verbs, which are mostly complex verbs, constructed with modal nouns/adjectives and a  light verb), directly or indirectly are loan words from Arabic.

Furthermore, modal  auxiliaries, and also some of the other modal elements, are polysemous, that is, they are  used to state several meanings in the domain of modality. In search for classifying  languages based on a semantic feature, besides the semantic map of modality in these  languages, two other methods were also suggested. One, classifying languages, according  to the number of the elements they apply to express modality. In this method, instead of  presenting branches which a language is or is not a member of, a continuum was proposed  where all languages laid on it based on the number of modal auxiliaries and modal  adverbs. The other, classifies languages, on the basis of their origin, where the modal  auxiliaries with the same source, are grouped together as one category.  

Modality and Evidentiality in Akan: A Corpus-Based Study - Joana Portia Saky (13/03/2019)

Joana Portia Saky

  • 13 maart 2019
  • Promotor: Jan Nuyts

Abstract

The existing literature has focused on the functional use of modality in terms of its qualification of propositions. The use of modal auxiliaries to express these qualifications has received quite an attention. This issue sometimes leads people to equate the study of modality to the study of modal auxiliaries only. In recent years, however, writers have shown that other word classes (adverbs, adjectives and verbs) are also used to express propositional qualifications in. As regards evidentiality, various studies have demonstrated the use of both grammatical and lexical paradigms for indicating the source of information. 

The current study is motivated by the lack of detailed discussion on these topics in the language and thus makes the effort to explore, describe and discuss the expression of modality and evidentiality by the use of available forms in Akan discourse. In line with the views writers have expressed on these topics, we demonstrate in this study that Akan uses modal auxiliaries, modal adverbs, predicative modal adjectives, verbs and verbal constructions to express epistemic, deontic and dynamic modalities. Further, we have pointed out that the language utilizes only lexical verbs, verbal and adjectival constructions to express evidentiality. 

On the relationship existing between modality and evidentiality, the study has proceeded on the common linguistic premise that everyone speaks based on a kind of evidence, knowledge or experience. Thus, whatever information one gives (be it stating a fact (non-modalized) or speculating (epistemically modalized) etc.) is based on evidence (see Nuyts 2001b; 1993). As such, forms and constructions in this study have been analyzed according to which modal or evidential meanings they express. In all instances, none of the forms identified and discussed in the study, expresses both modality and evidentiality. Of course, there are forms which can express two or more modal meanings or different degrees of modality in different contexts. Others too can express different forms of evidence but not both notions. 

Three of the Akan dialects, viz. Akuapem Twi, Asante Twi and Fante are involved in the discussions as a result of their literary status. From time to time, we indicate how a form or a construction is represented in the new Akan orthography we have renamed as Akanwa.

Metacognition in healthcare communication training: a critical analysis of blended second language learning - Christine Fourie (20/11/2018)

Christine Fourie

  • 20 november 2018
  • Promotor: Prof. Kris Van de Poel

Abstract

The diversity of learning needs among the members of any professional group of second language users is often under-estimated. By blending conventional classroom teaching with online learning, the linguistic needs of learners can seem scattered and beyond the teacher’s ability to address. However, for healthcare professionals, communication training in a second language is crucial, as sound communication between patient and healthcare provider is associated with improved patient health outcomes.

Consequently, my thesis wants to demonstrate how to optimise communication training for healthcare purposes via a blended learning platform by facilitating and supporting learner metacognition. The literature shows that attention spent on metacognition is known to improve learning outcomes independent of learner intelligence. Though often neglected in existing research, this thesis emphasises that metacognition only manifests when learners have the opportunity to exercise both declarative and procedural knowledge, which in a language learning environment translate as awareness of and reflection on learning a language (declarative knowledge) and the skill to implement this knowledge as engagement with learning via interaction (procedural knowledge). Apart from exploring learner metacognition, this thesis also explores the principles underlying the curriculum and syllabus design process as meta-metacognition, where needs analysis and evaluation can be defined as declarative insights that feed into implementation during the pedagogical phases of teaching and learning (procedural knowledge).

The research context of this thesis spans two continents, with the main studies conducted amongst (1) medical students at the University of Stellenbosch, South Africa, and (2) pre-professional nurses studying at Artesis-Plantijn Hogeschool in Antwerp, Belgium. To validate and compare results, data were also obtained from (3) a group of medical professionals in Gothenburg, Sweden, studying Swedish as a second language via an online platform, and (4) first and second year majors of English at the University of Antwerp, Belgium. The core syllabi for the medical as well as nursing training were based on the online modules Medics on the Move and Nursing on the Move, which were developed along the same foundations.

A multi-methods research approach was opted for in order to gain a comprehensive understanding of the metacognitive processes in the different blended learning contexts and the research approach taken was mainly qualitative as the research purpose was to investigate the nature and impact of metacognition in a healthcare community, rather than to test a theory or measure variables. The research designs of enquiry varied slightly between the different phases of the research. Quantitative data were collected via questionnaires containing Likert (type) scale items, for example, in the usability study adopted in a logframe model, as well as by tracking students’ online learning behaviour. Qualitative data were collected by means of transcribed focus group discussions, Facebook posts and open-ended questions and analysed according to the principles of grounded theory.

The results from the data analysis were organised in four parts to address the research question.

The first focal point of the research question is on the foundations of curriculum and syllabus design, where the first foundation entails a process of on-going needs analyses among stakeholders. This in turn should feed into the syllabus, as the next foundation. The syllabus should be functional, while relying on relevant scenarios to support communication. The third foundation refers to the task-based pedagogy underlying the course and realistically reflecting the activities to be undertaken in a healthcare context. Both teacher and learner should each take responsibility for learning during task performance. The fourth foundation then refers to stakeholder evaluations feeding continuously and systematically into the on-going development of the materials and teaching approach. Both the needs analyses and reflections function as declarative insights, which, when implemented, represent procedural knowledge. As the overall regulatory process, curriculum and syllabus design can be coined ‘meta-metacognition’, which will remain pro-active in delivering a relevant and meaningful learning experience. 

The second part of the research focuses on the dynamic nature of learning support that results from the meta-metacognitive processes (needs analyses and evaluation) and the reciprocal relationship between curriculum components and the stakeholders’ input (procedural knowledge).

The third part studies the nature of declarative and procedural knowledge, which stretches over three knowledge fields, namely (1) the self as a learner, (2) the task of learning a new language for healthcare purposes, and (3) strategies to support this learning process. Results indicate that collaborative learning in a social network fosters true peer interaction, while learners at the same time gain confidence in using the second language. Furthermore, the safe environment of the closed group makes it possible for learners to validate aspects of their identities as learners in a second language. Declarative knowledge (reflection) and procedural knowledge (interaction, validation) enable learners to take part in the on-going process of identity formation, which is necessary for their personal wellbeing. Results regarding the nature of language learning indicate that tasks should be interactive, relevant and personalised in order to support learning. Finally, results regarding learning strategies indicate that learning is ultimately an individualised process and that the blended learning approach offers learners a variety of strategies to choose from. The results furthermore indicate that there may be a discrepancy between a learner’s declarative and procedural knowledge, which can point towards unvoiced learner/learning needs.

Finally, based on a systematic review of all the analyses undertaken, this thesis produces a taxonomy of learner metacognition, a model of the meta-metacognitive processes and a set of guidelines to optimise healthcare communication training courses by facilitating meta-metacognition and learner-metacognition. In doing so, I hope to deliver a defensible curriculum and syllabus that can meet the diverse and fluctuating needs of the healthcare learner with a pro-active approach.

KEYWORDS: metacognition, blended learning, communication training, health-literacy, online learning, vocational training, collaborative learning, meta-metacognition

Accommodative competence: A Communication Accommodation Theory approach to language and communication training in nursing - Marilize Pretorius (20/11/2018)

Marilize Pretorius

  • 20 november 2018
  • Promotor: Prof. Kris Van de Poel

Abstract

Effective and appropriate communication in the nursing context is essential to the safety, care, treatment and well-being of patients. Language-discordance (i.e. where one or both interactants use an additional language) in the healthcare context is becoming nearly unavoidable due to the high mobility of both nursing professionals and patient populations. When language discordance impedes nurse-patient interaction, negative consequences can ensue. There is extensive descriptive research on the problems and consequences of language-discordant nurse-patient interactions in the healthcare, applied linguistic and communication literature, which has repeatedly emphasised the need for language and communication training for nursing (pre-)professionals. By comparison, research that puts forward and defends solutions to the problems outlined in the literature is scarce. Furthermore, the utility and validity of training programmes that have been reported on remain in question.

In this thesis I test a model for the design of language- and communication-training programmes for nursing (pre-)professionals based on communication accommodation theory. Communication accommodation theory can be conceptualised as a theory pertaining to competence in adjusting communicative behaviour to accommodate others. The purpose of accommodation is to manage comprehension, communicative efficiency, identity and social distance. Accommodating to patients in communication can be seen as an essential part of providing nursing care. Competently accommodating to patients when using an additional language to communicate, requires at least two types of accommodative resources: 1) communicative competence in an additional language, and 2) awareness and perspective-taking skills to enable nursing (pre-)professionals to gauge whether the use of their communicative behaviour in the additional language has been perceived as accommodative. This thesis focuses on the factors that may determine or influence whether nursing (pre-)professionals’ use of an additional language is perceived as accommodative.

The participants consisted of nursing (pre-)professionals working in multilingual, multicultural hospitals in South Africa, nursing students from Belgium undergoing English language and communication training, and internationally educated nurses, who speak English as an additional language, working in the United Kingdom. Various methodologies and tools were employed to investigate factors that affect perceptions and evaluations of the language-discordant nurse-patient interactions and the interactants as (non)accommodative

The factors identified can be grouped into two categories: 1) awareness of context and 2) perceptions of self and others. Awareness of context includes tapping into a) intergroup and b) interpersonal history. The former relates to being sensitive to intergroup identities, such as linguistic and cultural identities, as well as the roles that patients and nursing (pre-)professionals take on in the healthcare context. Interpersonal history, which pertains to the degree of nurse-patient rapport that already exists, can mitigate the effects of potentially negative associations with cultural or linguistic identities, and with healthcare-specific roles.

Nursing (pre-)professionals’ perceptions of patients’ affect and evaluations in language-discordant nurse-patient interactions, in comparison to their own affect and evaluations, motivates the nursing (pre-)professionals’ to develop and use strategic competence to solve communication problems caused by language discordance. Also, nursing (pre-)professionals tend to be unaware of their actual communicative competence; self-perceived communicative competence affects the manner in which they draw from their limited communicative competence to accommodate to their patients. However, raising nursing (pre-)professionals’ awareness about their actual communicative competence can lead to increased communicative confidence and self-efficacy beliefs; it can also increase the use of previously under-utilised communicative competence resources, which is the result of an under-estimation of competence. Also, gaining a more accurate perspective of their communicative competence can point nursing (pre-)professionals to their language and communication learning needs, while curtailing potentially noncaccommodative communicative behaviour resulting from an over-estimation of competence.

The salient factors and components have been incorporated into a model of accommodative competence for nursing (pre-)professionals. From the model, pedagogic implications and guidelines have also been derived for the design and implementation of training in accommodative competence. The training model focuses not only on communicative competence, but also on the awareness and perspective-taking skills that facilitates the strategic use of their communicative competence to accommodate to patients. In this way, the accommodative competence training model can be used to help nursing (pre-)professionals become more competent, confident and accommodative professionals.

Computational Mechanisms for Bootstrapping in Language Development: Discovering Categories in Speech - Robert Grimm (14/11/2018)

Robert Grimm

  • 14 november 2018
  • Promotoren: Prof. Waleter Daelemans en Prof. Steven Gillis

Abstract

When humans acquire the capacity to comprehend and produce language, they need to solve many interconnected tasks.  For example, they need to segment continuous speech into discrete units and assign meaning to those units. The available evidence indicates that knowledge acquired through progress on one task sustains and facilitates -- or bootstraps -- progress in other areas.  

Here, we construct computational models of bootstrapping processes in language development, which involve the usage of information from one domain in order to solve tasks from a different domain. Over the course of three studies, we first model how children use knowledge from the perceptual domain in order to discover linguistic units in unsegmented speech. This is followed by a final study that models how adult listeners use resources from the domain of normal hearing to break into category perception with cochlear implants.  

With respect to the former, we select multi-word and multi-syllable chunks from large corpora of child-directed speech, and we predict the age at which children first produce words based on the number of chunks containing each word. This approach assumes that if a particular word is contained in many unsegmented chunks, stored in children’s long-term memory, children should easily discover it, and they should begin to use it early in development.

We show that short syllable chunks, in contrast to frequent or internally predictable sequences, are best-suited for predicting the time course word learning. In addition, short sequences are also the most likely to correspond to words -- suggesting that children's early proto-lexica contain short, word-like chunks.  

Following the work on chunks, we describe a final study that simulates speech processing in adults with cochlear implants (CIs) -- neural prostheses used to partially replace a damaged inner ear. Many CI users transition from processing high-resolution signals, delivered through the inner ear, to lower-resolution signals delivered through the implant.

We model this in deep neural networks, focusing on interference between neighboring channels within CIs (channel interaction). We find that neural networks which are trained on high-resolution speech, prior to training on low-resolution data, learn more slowly if simulated channel interaction is present in low-resolution input.  The spectral degradation caused by channel interaction may thus require additional fine-tuning of existing neural circuits, slowing the transition to CIs after normal hearing.

The Networking Power of Web 2.0 in the Foreign Language Classroom: Affordances of the Online Peer Interaction Process - Ward Peeters (28/09/2018)

Ward Peeters

  • 28 september 2018
  • Promotor: Prof. Kris Van de Poel

Abstract

Since the use of Web 2.0 tools in language learning contexts has gained momentum, it is vital for practitioners to assess and evaluate the ways in which computer-mediated communication affects and constrains human interaction processes. At the same time, it is important to consider how users shape the online space itself, as well as how teachers’ and researchers’ understanding of these social semiotic spaces contributes to the development and refinement of current language learning activities. Fluctuating levels of success of online peer interaction projects suggest that we do not fully understand how learners build communities and develop their own agency within online environments. This study addresses one of the major gaps in present-day applied linguistic research: How do foreign language majors use the target language to learn, play and establish themselves as learners in a new learning environment when interacting with their peers online?

This study investigates the peer interaction process among first-year language majors studying English as a foreign language at the University of Antwerp in two consecutive years (Group 1: N = 119, Group 2: N = 112). A task-based approach, aimed to facilitate engagement and negotiation, was embedded in a Facebook environment, the most frequently used Web 2.0 tool among learners, which was then integrated into the undergraduate curriculum of English Proficiency. This resulted in a corpus of almost 6,000 online postings.

The threaded conversations were coded based on the principles of Grounded Theory and show how participants socialise and negotiate, as well as how the interaction process pertains to learning and to learners’ adaptation to the new academic environment. Learners’ engagement can be summarised in four developmental processes: 1) cognitive development, or making meaning; 2) metacognitive development, or reflecting on, synthesising and planning their learning and collaboration with others; 3) organisational development, or acculturating to the new academic environment; and 4) socio-affective engagement, or establishing and maintaining social cohesion. These four developmental processes form the basis of interpersonal communication and text creation in a social networking environment and exemplify how learners appropriate the online space for language learning purposes.  

Peer interaction and information exchange in the Web 2.0 setting –adapted to serve academic needs– supports learners in finding their own voice in their new educational environment as the project tapped into differentiation, appeal and rationality to design and implement an effective and sustainable peer interaction environment. This novel stance to providing online support in language learning contexts –where a task-based approach puts emphasis on learner activation, scaffolded support and co-dependence– enables learners to explore and co-learn with a community of peers. The project empowers them to observe others, test performance criteria and evaluate their own praxis, thus raising awareness of their own conduct and behaviour, as well as of the affordances and possibilities of peer interaction as online support.

This study succeeded in creating an online community of practice in which the individual was given the opportunity to grow and develop as a language learner while being part of a collective. It shows that formal assessment does not have to hamper the informal, interactive nature of Web 2.0, but that it can be an incentive for learners to start negotiating, socialising and learning online. This perspective on the affordances of the peer interaction process through Web 2.0 tools –where boundaries between formal and informal learning become blurred– provides learners with the opportunity to experience active language negotiation, and to orchestrate their own engagement with other learners and with the materials and resources available. This approach proves that adopting social networking tools to bridge between classroom instruction and learners’ out-of-class learning activities makes learning a social process of co-dependence, giving users ample opportunity to interact, make sense of their learning environment and exchange information to become fully-interactive, critical and autonomous language learners. 

Prosodic development in canonical babble and first words of cochlear implanted and normally hearing infants - Ilke De Clerck (24/09/2018)

Ilke De Clerck

  • 24 september 2018
  • Promotoren: Prof. dr Steven Gillis en Prof. dr Jo Verhoeven

Abstract

This dissertation aims to investigate the prosodic development in the canonical babble and first words of infants with cochlear implants (CI) and their normally hearing (NH) peers. Two main research questions are investigated:

  • When and how do NH infants who start to produce word stress?
  • When and how do infants with CI start to produce word stress?

These questions are investigated by five studies, with each different methodologies. In total two acoustic studies and three perceptual studies are conducted. These studies take a longitudinal approach, including data from the onset of babbling until a cumulative vocabulary of 200 words was reached, in both NH and CI infants. This range covers the period of the transition from canonical babbling to first words use. Within this range disyllabic babble and early word productions were selected for acoustic (Chapters 2 and 3) and perceptual analysis (Chapters 4, 5 and 6).

When and how do NH infants start to produce word stress? The results show that traces of the predominant trochaic stress pattern are already apparent from babbling onwards. This finding leads to the suggestion that infant babble serves as a training ground for ambient stress production. Moreover, the results show that the advent of first words boosts ambient prosody production. It is concluded that the communicative function of words and the fact that they have a clear adult target leads to refined stress production in typically developing infants.

When and how do infants with CI start to produce word stress? This dissertation uncovers a developmental delay in the production of the fine phonetic detail in CI infants. Children with CI only reach the level of prosodic differentiation in words NH children already display when they are babbling. Given the auditory deprivation and the initial delayed access to speech input, the idea of a developmental delay in prosody production is not surprising: children with CI had less aural exposure than their NH peers and they are still catching up at the time they enter the first word stage. The developmental lag that is already apparent in babbling evolves to an even bigger discrepancy in first word productions. At the time infants with CI start to produce lexical utterances they have had less time to exercise prosody production in canonical babble. This thesis is both clinically and scientifically relevant as it is amongst the first studies to detect such an early developmental delay.

Rationality and rhetoric.Higher education policy reform in Germany, 1992-2013 - Fergal Treanor (29/06/2018)

Fergal Treanor

  • 29 juni 2018
  • Promotor: Prof. dr Jef Verschueren

Abstract

When ideologies clash, conflict ensues and discursive change occurs. How do texts, often at odds with each other, work together to instantiate this discursive change? And how does discursive change make texts be the way they are? This dissertation analyses 165 original texts, compiled from a variety of sources, all of which participate – or seek to participate – in the formulation and implementation of policies relating to the degree structure reform (Studienstrukturreform) in Germany, 1992 – 2013. The texts are stylistically diverse, and represent a wide spectrum of differing opinions, but their common goal of policy formulation unites them in a Community of Practice (CoP) engaged in coordinative discourse. Analysis is corpus-driven and multifactoral, focuses on textual and transtextual structures, and directs particular attention to indexical values. Ten theses on transtextual phenomena are developed, the most important of which is the conceptualization of “rational” argument in discourse as a rhetorical performance, which orients recipients to preferred aspects of tacitly held knowledge schemata – or ideology – present in discourse.


Wanneer ideologieën botsen, volgt conflict en grijpen discursieve veranderingen plaats. Hoe werken teksten, die vaak onverenigbaar zijn, samen om dergelijke discursieve veranderingen tot stand te brengen? En hoe maakt discursieve verandering teksten tot wat ze zijn? Dit proefschrift analyseert 165 oorspronkelijke teksten, gecompileerd uit diverse bronnen, die allemaal deelnemen – of trachten deel te nemen – aan de formulering en implementering van beleid met betrekking tot de hervorming van de academische gradenstructuur (Studienstrukturreform) in Duitsland, van 1992 tot 2013. De teksten zijn stilistisch van uiteenlopende aard, en zij vertegenwoordigen een ruim spectrum van verschillende opinies, maar het gemeenschappelijk doel van beleidsvorming verenigt ze in een ‘Community of Practice’ (CoP) die volop betrokken is bij de ontwikkeling van gecoördineerd discours. De analyse is corpus-gestuurd en multifactorieel, concentreert zich op tekstuele en transtekstuele structuren, en vestigt de aandacht op indexicale waarden. Tien stellingen met betrekking tot transtekstuele fenomenen worden ontwikkeld. De belangrijkste is de conceptualisering van ‘rationele’ argumentatie in het discours als een retorische performance die de ontvangers stuurt naar gewenste aspecten van onuitgesproken kennisschema’s – of ideologie – aanwezig in het discours.

Two Authors Walk into a Bar: Studies in Author Profiling - Ben Verhoeven (27/06/2018)

Ben Verhoeven

  • 27 juni 2018
  • Promotor: Walter Daelemans
  • CLiPS

Abstract

This is a dissertation about author profiling in all its shapes and forms. We will identify three current needs in author profiling research that we deem important. These needs are addressed in several case studies that form the chapters of this dissertation.

Audio-description in Dutch: a corpus-based study into the linguistic features of a new, multimodal text type - Nina Reviers (05/02/2018)

Nina Reviers

  • 5 februari 2018
  • Promotoren: Prof. Aline Remael en Prof. Reinhild Vandekerckhove

Abstract

This PhD project is a corpus-based study of the linguistic features of a new, multimodal text type within Audiovisual Translation (AVT): Audio-description (AD) for the blind and visually impaired. The aim of this interdisciplinary project is to describe the lexico-grammatical features of AD-scripts and examine the role they play in the specific communicative function of the text. The object is to explore one of the key-issues in AD research: How are images put into words and what are the implications for the language use in AD? A recent pilot study confirmed the hypothesis that the language of AD contains distinctive grammatical (morpho-syntactic) and lexical features and that these specific patterns can be identified by corpus analysis. Firstly, the current project developed an extensive and varied text corpus of AD scripts of Dutch audio-described films and series. Secondly, this text corpus provided the basis for quantitative linguistic research, aiming to identify the prominent lexico-grammatical features of the text type. Thirdly, the quantitative analysis was combined with a qualitative analysis of the (communicative) function of these features.

Finally, special attention was paid to the multimodal nature of the text type, since the AD-script only makes sense in combination with the dialogues, music and sound effects of the original film or series with which it forms a coherent whole. Therefore, a multimodal analysis of a selection of texts was conducted. This multimodal analysis explored the unique interaction between the language of AD and the other channels of the audiovisual text, more specifically the sound effects.

Ultimately, the project’s ambition was to conduct an extensive linguistic audience design oriented analysis of the AD-discourse. This allowed us to identify the features that characterise the AD text type, clarify how these linguistic and stylistic features are used to ensure maximum communicative efficiency, and how these features are related to the function and multimodal character of AD.

The project presented here is a pioneer in the field: AD has become an international research topic recently but for Flanders and the Netherlands no study of AD is available yet. In addition, its results can offer the basis for future application-oriented studies and can support the development of a local AD tradition in Flanders that meets international quality standards.

Ideologies and Authoritarian Discourse in Wartime Japan 1931-1945: Metaphor and Discursive Markers in the media and soldiers' personal writings - Marie Reetz (01/12/2017)

Marie Reetz

  • 1 december 2017
  • Promotor: Prof. Jef Verschueren

Abstract

Under the rule of its Authoritarian and Militarist government during the 1930s until mid 1940s, Japan’s population is being mobilized as well as self-mobilizing through propaganda infused language. While Japan’s war of aggression is being portrayed as a “war of liberation” and independence led by a modern and progressive state, the Japanese are encouraged to suffer, fight or even die for the “Total War” their country wages throughout Asia.

My dissertation seeks to shed light on the specific language and discourse phenomena that seem to play a part in ideologically constructing language that naturalises (Althusser, Bourdieu) central themes of patriotism and war like “identity”, “sacrifice” and “death”. I understand the ideological mechanism to consist in presenting – in a socially and often politically meaningful way – a version of reality as THE reality.

Looking at government policies and soldiers’ writing, this work shows how language forms and constructions linked to the larger discourse context (Foucault’s archive) – metaphors, syntactic and semantic embeddedness, phraseologies (epistolary formulae) and genre – come together to create an impression of normalcy (effet d’évidence), naturalising Japaneseness and dying for Emperor and Country (okuni no tame).

Drawing on (French) Discourse Analysis, Critical Discourse Analysis and Pragmatics, I develop my methodology with insights gleaned from works on Fascist and Totalitarian discourse, on ideology and language use and from reflections on the parallels between ideology and (grammatical, conceptual and rhetorical) metaphors. The empirical chapters present quantitative distributional analyses and qualitative analyses of the media, school and epistolary corpora, containing the two main daily newspapers Asahi Shimbun and Yomiuri Shimbun, high-school ethics textbooks and soldiers’ diaries and letters.

How trilinguals process cognates and interlingual homographs: the effect of activation, decision, and cognitive control - Ihor Borisovitch Biloushchenko (16/10/2017)

Ihor Borisovitch Biloushchenko

  • 16 oktober 2017
  • Promotoren: Prof. dr Dominiek Sandra en Prof. dr Ton Dijkstra

Abstract

Het doel van dit onderzoek was om meer in detail te bestuderen hoe multilingualen interlinguale woorden verwerken die hun orthografie én semantiek over twee talen met elkaar delen (cognaten) of alleen hun orthografie (interlinguale homografen). In het bijzonder richten we ons op de rol van een aantal linguïstische en niet-linguïstische factoren die de herkenning van interlinguale woorden beïnvloeden, waaronder de stimuluslijstcontext, de taakvereisten en de dominantie van een taal. Daarnaast hebben we de cognitieve controlemechanismen bestudeerd die betrokken zijn bij de onderdrukking van de irrelevante betekenis van interlinguale homografen en hebben we deze mechanismen vergeleken met de mechanismen van cognitieve controle die gebruikt worden in niet-linguïstische taken.

In dit onderzoek staan een aantal vragen centraal. Hebben identieke cognaten meerdere representaties in het gedeelde meertalige lexicon? Wat is de invloed van verschillende linguïstische en niet-linguïstische factoren op de verwerking van cognaten en interlinguale homografen? Wat is het verband tussen de grootte van homograafinhibitie en de responssnelheid? Gebruiken proefpersonen vergelijkbare cognitieve controlemechanismen in linguïstische en niet-linguïstische taken? Al deze vragen werden onderzocht door in te zoomen op de bijdrage en combinatie van lexicale en post-lexicale processen tijdens het proces van meertalige woordherkenning.

De uitgevoerde experimenten benaderen de kwesties vanuit verschillende invalshoeken in termen van stimulusmateriaal, experimenteel design, onderzoekstechnieken, statistische analyses en proefpersonen. Een eerste belangrijke bijdrage van dit onderzoek betreft het gebruik van dezelfde set stimulusmateriaal in diverse experimenten, die slechts in geringe mate van elkaar verschilden in termen van hun opzet. De zorgvuldige, gedetailleerde manipulatie van onafhankelijke variabelen vergroot de kracht van statistische analyses over experimenten heen en zorgt voor een tot dusver niet benutte mogelijkheid tot theoretische vergelijking. Bovendien konden gedeelde kernkenmerken van het proces van meertalige woordherkenning onderscheiden worden van taakspecifieke componenten door de data in twee gerelateerde maar licht verschillende experimentele taken met elkaar te vergelijken (bv. lexicale-decisietaak en go/no-go-taak). Ten slotte, door drie talen met elkaar te vergelijken in plaats van twee (zoals in het verleden meestal het geval was) konden we de effecten van woordcompetitie van de dominantere taal met die van de zwakkere taal op de doeltaal met elkaar vergelijken. Daarnaast hebben we ook de focus gericht op de cognitieve controlemechanismen bij de verwerking van interlinguale homografen en die vergeleken met de controlemechanismen in niet-linguïstische taken om na te gaan of een verklaring van homograafinhibitie op basis van domeinonafhankelijke inhibitieprocessen mogelijk is.

Coding cognition: A case-study of Dutch-Turkish and Dutch-Moroccan Arabic language users in the Netherlands - Bram Vertommen (29/09/2017)

Bram Vertommen

  • 29 september 2017
  • Promotoren: Prof. dr Jef Verschueren en Prof. dr Karen Lahousse

Abstract

The information we share with other people has many different shapes. Some information describes or recalls experienced events. Other information refers to dreams, imaginations or intentions. Yet other information deals with how we try to make sense of the complex reality around us: e.g., tendencies or generalizations we can induce from recurrent things we experience, and predictions we can make about particular phenomena.

It is widely acknowledged that experience- or imagination-based reports involve different cognitive operations than generalizations that result from inductive reasoning. Moreover, this difference has been proven to be linguistically relevant, in that it can be signaled by specific linguistic elements.

In the dissertation I propose a heuristic that measures the impact of the conceptual distinction above – between information that contains direct perception (or imagination) based input and information that does not – on language choice in two bilingual varieties spoken by second-generation heritage speakers in the Netherlands (Nortier 1990; Backus 1996; Boumans 1998; Eversteijn 2011).

These varieties are characterized by intensive and consistent alternations from a grammatical structure defined by one language (e.g., Dutch) to a grammatical structure defined by the other (e.g., Turkish or Moroccan Arabic), and vice versa. Through my heuristic I intend to demonstrate that, unlike many alternations previously investigated in multilingual studies (Auer 1998), the ones in the data sets under scrutiny do not tend to signal local extralinguistic or interactional aspects (e.g., a change of participant constellation, a change of activity type) Rather, I argue that sentences are in one or the other language because crucially different cognitive operations underlie the information expressed by these sentences.

Central in the heuristic is the logical distinction between categorical judgments (e.g., Manon is a dancer) and thetic judgments (e.g., (What is happening?) – Manon is dancing on the lawn) (Kuroda 1972). Contrary to the common view that both judgments are difficult to empirically verify (Sasse 1996), I argue that these notions can be operationalized if closely related and syntactically reflected conceptual distinctions are taken into account. First, categorical judgments tend to depict fundamentally different types of states-of-affairs (i.e., general states) than thetic judgments do (i.e., transitory events). Second, I claim that the states-of-affairs contained in categorical and thetic judgments are subject to different subtypes of quantification: categorical judgments depict universally quantified states-of-affairs whereas thetic judgments deal with existentially quantified states-of-affairs. A universally quantified state-of-affairs is construed as being true under well-defined general conditions. In an existentially quantified state-of-affairs one or multiple manifestations of this state-of-affairs are pointed at.  

The results of the corpus-based quantitative analysis strongly suggest that the formal variation between Dutch and Turkish or Moroccan Arabic in the bilingual data is paired with a conceptual distinction between categorical and thetic judgments (as defined above). This particular pairing of form and meaning also provides evidence for the fact that language choice, and the alternation between two languages, can be considered an integral part of the ‘grammar’ of the bilingual varieties used in the analyzed data sets.

References

  • Auer, Peter (ed.). 1998. Code-switching in conversation: language, interaction and identity. London: Routledge.
  • Backus, Ad. 1996. Two in one: bilingual speech of Turkish immigrants in the Netherlands. Tilburg: Tilburg University Press.
  • Boumans, Louis. 1998. The syntax of codeswitching: analysing Moroccan Arabic/Dutch conversations. Tilburg: Tilburg University Press.
  • Eversteijn, Nadia. 2011. “All at once”: language choice and codeswitching by Turkish-Dutch teenagers. Saarbrücken: Lap Lambert Academic Publishing AG & Co. KG.
  • Kuroda, Sige-Yuki. 1972. The categorical and the thetic judgment: evidence from Japanese syntax. Foundations of language 9(2). 153-185.
  • Nortier, Jacomine. 1990. Dutch-Moroccan Arabic code switching among Moroccans in the Netherlands. Dordrecht: Foris Publications.
  • Sasse, Hans-Jürgen. 1996. Theticity. Arbeitspaper Nr. 27 (Neue Folge). Köln: Institut für Sprachwissenschaft, Universität zu Köln.

The Pragmatics of Translation in Journalism: an Investigation into the Nature of Translation in the News Room - Maarten Charles J. Franck (03/07/2017)

Maarten Charles J. Franck

  • 3 juli 2017
  • Promotoren: Prof. dr Jef Verschueren en Prof. dr Leona Van Vaerenbergh

Abstract

This book explores the nature of translation in the newsroom from a linguistic pragmatic perspective. I have defined pragmatics as the study of communicative dynamics. Because of how it has been institutionalized, it is mainly a “general functional perspective on (any aspect of) language” (Verschueren 1999, p. 11). Looking at translation from said perspective, means that it can be regarded as a form of interpretive language use which is always dependent of context. It is an instance of recontextualization, in which a source text (ST) is manipulated to become a target text (TT). It can be interlingual or intralingual, depending on whether translation is done between two different general languages or within one general language. And while a translation is never completely faithful or liberal, it is always situated on the cline of unremitting variability between these two extremes. Translation is also one of many tasks associated with the profession of journalism. The main commodity of this profession is information. The information journalists provide can be true, but it is not always possible to say whether it is or is not, because most often, it is the neutral rendition of what other people have said.

With these basic notions in mind I set out to answer four main questions: (i) Which variables influence the way journalists translate?; (ii) What formal shifts in meaning occur when journalists translate?; (iii) What functional shifts in meaning occur when journalists translate?; (iv) Are metamessages strengthened throughout subsequent translations? To answer these questions I examined translations made by journalists working for Belgium’s main national press agency (Belga), news sites (www.demorgen.be, www.destandaard.be, www.hln.be and www.nieuwsblad.be), and newspapers (De Morgen, De Standaard, Het Laatste Nieuws and Het Nieuwsblad). The only one of these media that did not exclusively translate into Dutch was press agency Belga, which also translates news into French. Belga also often fulfilled an intermediary role, translating international media reports (e.g. Agence France-Press, Deutsche Presse Agentur) into French and Dutch before they were picked up by Belgian media. It also provided original coverage which was often picked up by the different media.

To account for differences in topic I exemplified two distinct cases in this book: (i) translations of news reports on the 2011 elections in the DR Congo; (ii) translations of news reports on the run-up to the 2012 London Summer Olympics.

Maritime English put to the test! The feasibility and desirability of setting global standards for Maritime English: a survey-based study - Allison Noble (29/06/2017)

Allison Noble

  • 29 juni 2017

Leden van de doctoraatscommissie

  • Prof.dr. Tom Van Hout, University of Antwerp, België (Chair)
  • Prof.dr. Lieve Vangehuchten, University of Antwerp, België
  • Prof. Willy Van Parys, Antwerp Maritime Academy, België
  • Prof.dr. Boris Pritchard, ret’d. University of Rijeka, Kroatië
  • Prof. Clive Cole, ret’d. World Maritime University, Zweden
  • Prof.dr. Dieter Vermandere, University of Antwerp, België

Les transformations de sens dans les médias. Une étude de la représentation du discours autre dans la presse écrite congolaise, française et belge. Le cas du conflit au Nord-Kivu en 2008 - Nancy Decloux (22/06/2017)

Nancy Decloux

  • 22 juni 2017
  • Promotor: Prof. dr Patrick Dendale

Abstract

Il est largement répandu que les médias transmettent une certaine image de la société. Cette image est déterminée par une série de facteurs culturels, comme les tendances politiques d’un pays, les institutions, les lignes éditoriales, etc. Dans le contexte de la mondialisation, la représentation de faits d’actualité étrangers mène à une possible comparaison entre ce que l’on pourrait appeler « une presse globale » et « une presse locale », la première référant aux fournisseurs médiatiques internationaux (occidentaux) et la seconde à une presse nationale, dans notre cas la presse africaine. Dans ces deux types de presse, la représentation d’un événement implique des différences ou des similarités dans la manière dont est utilisé le langage pour présenter des acteurs, des actions, des énonciateurs et des discours.

Dans cette thèse, nous avons exploré les transformations de sens lorsqu’un événement africain était non seulement relaté dans la presse africaine, mais aussi dans la presse internationale. Analysant un corpus de huit journaux francophones composés de textes sur le conflit qui a eu lieu en 2008 au Nord-Kivu en République démocratique du Congo (RDC), nous avons identifié les transformations de sens en Belgique (Le Soir, La Libre Belgique), en France (Le Monde, Le Figaro, Libération) et en RDC (Le Potentiel, L’Observateur, Le Phare). Concentrée sur la représentation du discours autre, nous avons montré que : (i) l’idée selon laquelle l’information circule du pays dans lequel a lieu l’événement vers des pays étrangers est contredite par nos analyses. (ii) Cet accès à l’information mène les journalistes congolais à utiliser des mécanismes énonciatifs pour montrer leur accord ou leur désaccord avec les journalistes ou avec les discours représentés dans les textes occidentaux. Si la représentation de la société dépend de la représentation des discours, la catégorisation des énonciateurs et des actes de parole représentés révèle les points de vue sous-jacents de chaque journal étudié. Pour comprendre ces points de vue, plusieurs analyses ont été menées : (a) la caractérisation des énonciateurs représentés qui jouent soit le rôle de journalistes (ex. : « Colette Braeckman dit… » ; « Selon Le Soir… »), soit le rôle d’acteurs du conflit représenté (ex. : « Les soldats ont dit… ») (b) La catégorisation des actes de parole représentés, par laquelle nous montrons que l’accord ou le désaccord avec les discours des énonciateurs représentés dépend aussi des actes de parole représentés. Plusieurs valeurs comme le doute/la certitude, la vérité/la fausseté, les effets sur les lecteurs, les relations à d’autres discours sont alors examinées. (c) Les formes de représentation du discours autre (les discours direct et indirect, les modalisations, etc.). Pour tous ces types d’analyses, nous nous sommes concentrée non seulement sur les types de transformations, mais aussi sur leurs fonctions (positionnement, réception, doute, certitude, etc.). Pour interpréter ces fonctions, une grande partie de notre travail s’est appuyé sur les notions de dialogisme et d’intertextualité.

Avec cette étude, nous avons voulu mieux comprendre l’accès à l’information en RDC, mais aussi comment les aspects de l’idéologie pouvaient affecter la manière dont les discours sont présentés et représentés dans la presse locale et la presse globale.

Writing in a third language A study of upper secondary students' texts, writing processes and metacognition - Yvonne Knospe (02/06/2017)

Yvonne Knospe

  • 2 juni 2017
  • Promotoren: Prof. Luuk Van Waes, Prof. Mariëlle Leijten en Prof. Kirk Sullivan
  • Dubbeldoctoraat UAntwerpen - Umeå University (Zweden)

Abstract

A Study of Upper Secondary Students’ Texts, Writing Processes and Metacognition Learning an additional foreign language (usually referred to as a third language, L3) after English (L2) in formal settings seems to remain relatively unsuccessful in the European context (European Commission 2012), despite the reported advantages from extensive language learning experiences. Against this background the present thesis explores the potential benefits of a teaching approach focused on writing strategies and metacognitive reflections for L3 writing.

The study investigates the influence of an intervention concerning the writing of argumentative texts in L3 German. Two classes with 15- and 16-year-olds from a Swedish upper secondary school participated. One class received writing instruction, the other one no special treatment. Texts were collected at three points in time. Additionally, seven students from the intervention class, the “focus students”, attended five individual writing sessions, in which their argumentative writing was recorded with keystroke logging and screen-recording software. After each writing session, individual stimulated-recall interviews were conducted.

The development of L3 writing throughout the course of the intervention was investigated from three different perspectives: writing products, writing processes and metacognitive knowledge. Text quality was studied on the basis of a selection of class texts, which were scored both analytically and holistically. The writing processes were examined by automated and manual analysis of the conducted log- and screen-recording files. Metacognitive knowledge was analysed based on transcripts of stimulated-recall interviews.

The findings show that only the texts of those students, who attended both the intervention and the individual writing sessions, including reflective interviews, improved in quality during the intervention. Further, these students slightly increased their writing fluency through higher text production rate and a shift from intense online-source use to more revisions. The investigation of metacognitive knowledge revealed a number of learner- and language-related variables, which seem to be decisive in understanding the interplay between learners’ attitudes, performance and strategic behaviour.

The results suggest that writing practice is crucial, but learning to write can be further enhanced by including writing strategy instruction and metacognitive reflections in the foreign language classroom.

Speech production and speech production accuracy in young children: hearing and hearing impaired children with a cochlear implant - Jolien Faes (03/05/2017)

Jolien Faes

  • 3 mei 2017
  • Promotor: Steven Gillis

Abstract

This dissertation considers the longitudinal phonological development of children with cochlear implants (CI) in comparison to that of normally hearing peers. The spontaneous speech productions of children with CI are investigated both at the word level for accuracy and variability and at the segmental level with respect to consonant clusters and fricatives. Analyses showed that children with CI’s speech production improves with age.

Words and segments become less variable and more accurate. Slightly better outcomes are found in children that receive their implant at an earlier age. In comparison to normally hearing peers, children with CI’s word productions are less accurate and more variable. Moreover, the characteristics of the target word (syllable length, complexity and frequency in child-direct speech) affect children with CI’s word productions to a different extent.

Target words that have more syllables and are more complex are less accurate and more variable in both groups of children, but the effect is more pronounced in children with CI. In contrast, target words that are more frequent in child-directed speech are more accurate, but this effect is less pronounced in children with CI. At the segmental level, children with CI’s consonant cluster production is delayed as compared to that of normally hearing age-mates.

However, for all of these measures, a catch up is found: children with CI reach similar levels of phonological development as their normally hearing age-mates by age five for accuracy and variability at the word level and for consonant cluster production. With respect to fricatives, results showed that children with CI are delayed in their production when compared to normally hearing age-mates, but not when compared to normally hearing children with an equal lexicon size. In other words, phonological development and lexical development are commensurate in children with CI, similarly to NH children. Next to phonological development, sentence length and inflectional richness were investigated.

Results showed that normally hearing children produce longer sentences and have a higher inflectional diversity than children with CI. Similarly to the results considering phonological development, children with CI are found to catch up on their normally hearing peers: by age five for inflectional richness and by age seven for sentence length. Thus, the present dissertation showed that it is crucial to trace the long-term language development of children with CI’s spontaneous speech productions.