Programm 71. StuTS + 31. TaCoS

Breakfast @ accommodation

Opening Plenary

More than words: Recognizing speech of people with Parkinson’s disease

ARAY: Create your own speaking clock

In this workshop we will briefly introduce speech synthesis, also known as text-to-speech (TTS), and focus on one particular type: concatenative speech synthesis. Then we will turn to code and show you an example: ARAY — world's first demotivation...

Crash Course Historical Comparative Method, The Tree Model & Genetic Relationship

We all know language families and language trees which are supposed to show how languages are related. But the method that underlies such trees, the theoretical background and its problems are not taught in my linguistic programms. Here's a worksh...

Language families and language trees do not fall from the sky. They are based on a theoretical framework called the (historical) comparative method. This method was developed in the 19th century for the Indo-European languages and assumes that languages split from earlier languages in isolation and evolve in systematic ways that lead to regular correspondences in the area of phonology and also morphology. The method does not take borrowings, linguistic diffusions and other phenomena of language contact into account and it remains unclear to which extent it can be applied to languages in general.
Unfortunately, the method and problems that come with language trees are not taught in many synchronically orientated linguistic programs so that students have to rely on and believe in existing trees. However, synchronic linguists should now about the theoretical background in order to evaluate trees and distinguish typological and genetically based similarities between languages. This is especially relevant for linguists working in the field who want to detect influences from neighbouring languages or people involved in synchronic comparative work. In this crash course we will take a...

Creating Dialect Maps (with QGIS)

In this workshop, we will go through:

What are… & how do I make…:
- Symbol maps
- Frequency maps
- Isogloss maps
- Area maps (we won’t go through this with QGIS due to time)

In dialectology, the study of geographical variation of language, we often encounter a huge amount of data of speech varieties from different places. In order to visualize the linguistic data of this kind, we often have to plot them on maps.

Dialectology as a separate discipline is rarely taught on its own; it is usually taught as a minor area in Sociolinguistics. Because of that, students don’t have a lot of exposure to the varieties of dialect maps out there, and more importantly, how to produce them on their own. In this workshop, I aim to fill the gap by firstly introducing several types of dialect maps, namely point maps, frequency maps, isogloss maps and area maps. The second part of the workshop will be on creating digital dialect maps using QGIS.

The mapping skills you learn in this workshop is not only for dialectology, they are also applicable for other areas and disciplines too, so be sure not to miss out!

*******IMPORTANT*******
Please download QGIS before the start of the workshop here:
https://qgis.org/en/site/forusers/download.html

In order to follow the mapping tasks during the workshop, please also download the files here beforehand:
https://dr...

Frisian Workshop

This workshop is an introduction to Frisian, the language spoken in the area where the conference is taking place this year. We will start with a short general introduction about Frisian: where it's spoken, its history and its status as a minority...

The rest is syntax

I present preliminary research on compositional phrase embeddings using a novel tree LSTM autoencoder. I will outline the necessary theoretical background for grammar-theoretical, neurolinguistic and NLP notions of (non-)compositionality and intro...

In this short talk I present my preliminary work on surprisal measures for text reuse retrieval.

One problem that deliniates (literary) text reuse from plagiarism is the use of wordplay and meaning altering substitutions on the quoted phrase. Arthur Golding, for example, writes "Instead of legs, to both her sides stick fingers long and fine: The rest is belly.” when describing the fate of Arachne in his translation of Ovid’s Metamorphoses (1567), which is later invoked in the dying words of Hamlet (”The rest is silence.”). Shakespeare is referencing Ovid (through Golding) to get to give his tragedy a mythological, heroic vibe. To do this, he is performing a substitution from belly to silence to match his scene.

To put it in concrete terms, there is a transposition from the global intertextual dialogic context the phrase is retrieved from to the local context of the borrowing text. My hypothesis is that the resulting violation of the global context should be surprising.

Surprisal here can be thought of in the information-theoretical sense of information content. If these borrowed phrases are lexically embedded “as is”, we expect with high probability to see the phrase u...

Is text length determined by register or by elicitation order?

One easily noticeable feature of register variation is the use of ellipses, reduced forms and deletions. The aforementioned are more prominent in informal registers (Biber, 1995, p. 4, Biber and Conrad, 2009, p. 79, p. 194). If the same contents were to be produced
in different registers, this should lead to shorter texts in informal registers in comparison to more formal ones. Since there seldom is the need to express the same contents in different registers, this topic is not easy to research.
The RUEG corpus (Wiese et al., 2021) contains four texts per participant in two different registers and two different modes: formal written, formal spoken, informal written, and informal spoken. The texts all contain descriptions of a minor car accident previously
shown to participants in a short video clip. Therefore, the RUEG corpus provides the opportunity to investigate such topics. There was no minimum or maximum text length requirement to be met by the participants during elicitation. To counterbalance effects stemming from elicitation order of the four texts, the order was randomised. Without such measures, an effect could arise in the form of decreasing text lengths from the f...

German loanwords in Cameroonian Fàŋ-Bətí languages

Basically, this study is geared towards the phonological description of the nouns the Fàŋ-Bətí languages borrowed from the German language. During the analysis, references would be made to other formal linguistic aspects of these loanword, e.g., t...

Basically, this study is geared towards the phonological description of the nouns the Fàŋ-Bətí languages borrowed from the German language. During the analysis, references would be made to other formal linguistic aspects of these loanword, e.g., the semantic, and morphological aspects, only where necessary. One of the methodologies will be separate the words that have been borrow from English from those which have been borrow from German.

Traces of Germany’s colonial presence in Cameroon between 1884–1916, exist in the number and quality of words which Cameroonian indigenous languages, e.g. the Fàŋ-Bətí languages, borrowed from the German language. The German loanwords are one of the features found in various languages in the German formal colonial countries, eg Cameroon. The acquisition of these can be justified by variant degrees of language contact in the contexts of trading, schooling, missionary activities, and administration. The result of this language contact includes the development of full-fledged new language varieties, e.g. Pidgin-English (Mühlhäusler 1977) and creole language (Rabaul Creole German/Unser deutsch und Volker 1991, Maitz & Volker 2017), as well the ...

Fine-Tuned Sentence Transformer Model for Question Answering Task

I would like to introduce the structure and a practical application of sentence transformer models which are used for sentence- and document-level NLP tasks, such as information retrieval, text classification and question answering.

Nowadays, various pre-trained language models based on representation learning are constantly being developed and have achieved splendid performance in the application of various natural language processing (NLP) tasks by fine-tuning. Sentence Transformer models provide with the possibility to embed sentences and compare the semantic similarity between sentences by creating a siamese and triplet networks based on pre-trained Transformer models. The paper fine-tunes a Sentence Transformer model and applies it to the question answering (QA) task by answer selection. To be more specific, we use the finetuned sentence transformer model to select the proper answers to a given question from a given pool of candidate answers. The experiment results show that by fine-tuning, the accuracy can be improved significantly from 0.2664 to 0.4867. Further research should be done on fine-tuning more models and training on more domain data.

On Forced-Alignment and its Utility in Corpora for Phonetic Research

The VoxClamantis Corpus is a large corpus containing recordings of Bible readings, which have been forced aligned. The creators of the corpus promote it as a tool for large-scale typological research in phonetics. In this talk I first give some background on forced alignment and how it was applied to this corpus. I then I describe a small study I carried out which aimed to use the corpus to investigate which acoustic correlates differentiate the voiced aspirated affricate series from the voiced unaspirated affricate series in Bengali and Marathi. I discuss the benefits and limitations of using a forced aligned corpus for this type of research. Finally, I conclude that more work needs to be done to improve the performance of forced aligners before corpora such as the VoxClamantis Corpus can be used reliably for phonetics research, both for small-scale studies like mine, and for the large-scale typological research that the creators of the corpus intended.

"Kachten aus dem Kaatenspiel"

Eine laufende Untersuchung über die Verwendung des stimmlosen uvularen Frikativs als /r/-Variante.

Die Voreifel gilt als noch recht dialektstarke Gegend, der Dialektabbau ist hier im Vergleich zu anderen Regionen des Rheinlands am wenigsten fortgeschritten (vgl. Cornelissen 2005). Zwar geht auch dort die Zahl der Dialektsprechenden immer weiter zurück, jedoch verfügen viele über passive Dialektkenntnisse, die wiederum Einfluss auf die Regionalsprache ausüben (vgl. Cornelissen/Stiehl 2002). Bisher wurden zur Untersuchung solcher regionalsprachlichen Merkmale meist NORMs (Non-mobile Old Rural Male Speakers) befragt (vgl. Macha 1991). Im Gegensatz dazu konzentriert sich dieses Forschungsvorhaben auf Schüler*innen. Auch sie sind in der Regel am Erhebungsort aufgewachsen (und damit non-mobile), wodurch überregionale sprachliche Einflüsse als weniger starke Einflussfaktoren gewertet werden können. In der Schule werden Schüler*innen jedoch mit einem normorientierten Standard konfrontiert, die Kernlehrpläne weisen ‚die Standardsprache‘ als einzig akzeptable Unterrichtssprache aus (vgl. KLP 2013). So ergibt sich ein Spannungsverhältnis zwischen der überregional ausgerichteten Standardsprache und der regionaltypischen Sprechweise (vgl. Lameli 2011).
Dieses Spannungsverhältnis soll exe...

Wie Menschen in laufende Gespräche einsteigen

Wie gelingt es Menschen, erfolgreich in bereits laufende Interaktionen einzusteigen? Diese Frage wurde zuvor kaum linguistisch untersucht, auch wenn die theoretische Grundlage seit 1979 in Goffmans Participation framework besteht. Dieses erfasst a...

Der Vortrag beschäftigt sich mit einem Thema, das vielen aus dem Alltag bekannt sein dürfte: Wie steigen Menschen, die zuvor nur passiv zugehört haben, in eine bereits laufende Interaktion ein? Solche Szenarien sind zwar in der Realität häufig anzutreffen, passen aber nicht recht in das traditionelle Schema aus ‚Sprecher‘ und ‚Hörer‘. Daher wurde für die vorgestellte Untersuchung das Participation framework von Erving Goffman (1979) zugrunde gelegt, das auch die sozialen Kontexte von Interaktionen in den Blick nimmt, in denen auch andere, nicht ratifizierte Anwesende als Overhearer zuhören und in Gespräche einsteigen können. Solche Gesprächseinstiege wurden allerdings noch nicht empirisch untersucht, im Gegensatz zu Turnübernahmen ratifizierter Interagierender (Ford 2008, Mondada 2012).
Die vorgestellte Untersuchung hat daher zum Ziel, anhand authentischer Videodaten sprachliche Formen zu identifizieren, die bisherige Overhearer für die Funktion des Gesprächseinstiegs einsetzen. Die Daten stammen aus einem Friseursalon mit zwei nebeneinanderliegenden Schneideplätzen, sodass das Setting durchaus Voraussetzungen für Gesprächseinstiege bietet; es handelt sich aber um eine natürlic...

R-Sandhi: Do learners of EFL shadow native speakers?

As part of a sociophonetics seminar by Prof. Sabine Arndt-Lappe (Uni Trier) an experiment was designed to find out, whether learners of English as a Foreign Language shadow native speakers.

Shadowing experiments are generally used to find out how long it takes for a person to process and produce speech, but they can also be designed to elicit other data. As part of a seminar on sociophonetics by Prof. Sabine Arndt-Lappe (Trier University) an experiment was designed to find out whether learners of English as a Foreign Language shadow native speakers. A remote experiment was conducted by students. Participants were asked to record themselves reading out sentences to establish a baseline. Afterwards participants recorded themselves shadowing audio recordings of a native speaker of British English. Together the recorded data was annotated for realisation of r-sandhi in the baseline recording and the shadowing task. The results of this experiment will be presented in this talk.

An analysis of strategies and tactics used in courtroom dialogues

Language used in courtrooms has long been an issue of interest, due to its strict formality and the power imbalance that exists between attorneys and witnesses. Yes/No Interrogatives (shortened as YNIs) are used overwhelmingly when lawyers cross-e...

References
Jefferson, G. (2004). Glossary of transcript symbols with an introduction. In G. H. Lerner (Ed.), Conversation
Analysis (pp. 13–31). John Benjamins Publishing Company.
Mortensen, S. S. (2020). A question of control? Forms and functions of courtroom questioning in two different adversarial trial systems. Scandinavian Studies in Language, 11(1), 239-278.
Raymond, G. (2003). Grammar and social organization: yes/no interrogatives and the structure of bonding. American Sociological Review, 68(6), 939-967.

Von Sprachkontakt und Zweckentfremdung: die Entwicklung des baskischen definiten Artikels

Die morphologische Markierung von Definitheit mittels Artikeln ist häufig ein areales Merkmal, das auch von genetisch nicht verwandten Sprachen innerhalb eines Gebiets geteilt wird. Ein solches Areal stellen Westeuropa und der Mittelmeerraum dar, wo neben verschiedenen indogermanischen und semitischen Sprachen auch Baskisch als isolierte Sprache einen definiten, ebenso wie einen indefiniten, Artikel besitzt. Der definite Artikel ist stärker grammatikalisiert, er wird an das letzte Element der Nominalphrase suffigiert und hat mehrere kasusabhängige Allomorphe. Auffällig ist, dass er nur in wenigen Kontexten tatsächlich auf definite NPs beschränkt ist. In allen anderen Kontexten tritt er auch mit non-spezifischen (nicht aber mit spezifischen!) indefiniten NPs auf, ebenso findet man ihn an prädikativen Adjektiven. Diese Verwendung des baskischen definiten Artikels ist semantisch nur begrenzt erklärbar und weicht auch von dem üblichen Grammatikalisierungspfad für definite Artikel ab. Tatsächlich ist der Gebrauch eher morphosyntaktisch als semantisch motiviert und der Verlauf der Grammatikalisierung wurde sowohl von sprachinternen morphosyntaktischen Faktoren als auch von romanischer...

"Warte, warst du jetzt im Krankenhaus?!"

Aus Imperativen entstandene Diskurspartikeln sind bisher vorwiegend in medial mündlicher Interaktion untersucht worden. Die Arbeiten zu komm (vgl. Proske 2014) und guck mal (vgl. Günthner 2017) sind zwei Beispiele für die Untersuchung dieser beson...

Diskurspartikeln waren bisher vorwiegend Untersuchungsgegenstand der Gesprochene-Sprache-Forschung und der Gesprächsanalyse (vgl. Proske 2014: 122), gelten sie doch als Charakteristika der konzeptionellen Mündlichkeit. Konzeptionell mündliche Elemente finden sich jedoch auch in interaktionaler computervermittelter Sprache. Insbesondere Diskurspartikeln, die aus Imperativen entstanden sind, wurden noch nicht umfassend in medial schriftlicher Inter-aktion untersucht.
Die Diskurspartikel warte ist in der Chat-Interaktion von besonderem Interesse, da zwischen den Interagierenden eine räumliche Trennung besteht, die eine wörtliche Interpretation des Gesagten ausschließt (vgl. dazu Günthner 2017: 17). Wenn Person A Person B in einer Textnach-richt mitteilt, dass sie warten soll, ist (nicht immer) gemeint, dass Person B ihre aktuelle Tätigkeit unterbrechen oder keine Antwort verfassen soll, bis Person A sich erneut gemeldet hat. Ähnlich wie in Face-to-Face-Interaktion erfüllt die Diskurspartikel noch weitere Diskursfunktionen, wie das Anzeigen eines Missverständnisses. In welchen Kontexten die Diskurspartikel auftaucht und welche Funktion sie einnehmen kann, unterscheidet sich jedoch...

Effects of Impoverishment and Obliteration in Morphology

In my talk I will lay out the different effects of the morphological operation Impoverishment and Obliteration. I will share insights from theoretical approaches to Impoverishment and Obliterations that have been introduced in the framework of Dis...

I will focus on the effects that both operations have been assumed to account for.
One such approach made by Harley (2008) is used to account for a meta-syncretism in Latin dative and ablative plural case exponents.
I will continue talking about her approach to Baoan case syncretisms, a peculiar case where Impoverishment offers an elegant way of solving a problem of high effort.
Next I will try to lay out Arregi & Nevins (2012) approaches to Participant Dissimilation and show how both Impoverishment and Obliteration are used here to account for these effects and also how the parallelism of these operations gives rise to an elegant solution of dialectal variation.

References:
Bonet, Eulalia (1991): Morphology after syntax: Pronominal clitics in Romance. MIT.
=> Often credited as the first proposal of Impoverishment
Harley, Heidi (2008): When is a syncretism more than a syncretism? Impoverishment, Metasyncretism, and underspecification. Phi-theory: Phi-Features across modules and interfaces pp.251-294.
Arregi, Karlos and Andrew Nevins (2012): Morphotactics: Basque auxiliaries and the structure of spellout. Vol. 86, Springer Science & Business Media.
=> Extensive...

Genesis and Typology of Mixed Languages

Game Night @ Accommodation

Breakfast @ accommodation

Accent stereotypes and their impact on language processing

People tend to categorize others based on their ethnic background, regional membership, and/or social class. We do not even have to see the other person for stereotypes to be activated: even a simple “hello” can be enough for a listener to sociall...

Intercultural Communicative Competence in foreign language teaching and learning: A comparison of the situation of Morocco and Germany in tertiary education

My participation aims at giving insightful criteria in choosing the research methodology, data collection tools and procedures; and identifying data analysis steps. The objective of my presentation is also to shed light on the importance of conduc...

Modern communication increases interaction through the use of new technologies which, in turn, urges learners to communicate with people who are geographically distant, and linguistically and culturally different. Priority is given to the educational system to prepare learners to the 21st century intercultural encounters and global interdependence. Linguistics, in general, and some of its ramifications, in particular, such as sociolinguistics and intercultural pragmatics have highlighted the intertwined relationship between language and culture (Risager, 2007). Foreign language education is itself, by definition, an intercultural enterprise (Secru, L et al, 2005). My empirical research aims to investigate how intercultural communication (IC) is taught in the English departments at Moulay Ismail University (Meknes, Morocco) and Bonn University (Bonn, Germany). It also identifies the EFL teaching strategies and their impact on students’ performance. More specifically, this research seeks (a) to test the level of intercultural communicative competence (ICC) of students enrolled in these two universities, (b) to compare the level of ICC among students who are affiliated in intercul...

Sound symbolism im cross-linguistischen Vergleich

In meiner Bachelorarbeit habe ich mich mit dem Thema sound symbolism in der Produktion beschäftigt. Die konkrete Frage, die ich mir gestellt habe, lautet: Verwenden Sprecher verschiedener Sprachen sound symbolism gleich oder gibt es Unterschiede, die von der Phonologie und / oder Phonotaktik ihrer Muttersprache abhängen? Um das zu untersuchen, führte ich ein Experiment mit (einer sehr kleinen Gruppe von) Muttersprachlern des Spanischen und des Polnischen durch, die gebeten wurden, neue Namen für Fabelwesen zu erfinden, die aus Köhlers (1933) „Maluma“- und „Takete“-Figuren entwickelt wurden. Ich habe mich für Polnisch und Spanisch entschieden, weil sich diese Sprachen in ihrer Phonologie stark voneinander unterscheiden: Während Polnisch über ein großes Konsonanteninventar verfügt, lange Konsonantencluster und komplexe Silben verwenden kann, hat Spanisch ein eher kleines Konsonanteninventar, kann verzweigende Nuklei verwenden und bevorzugt einfache Silben. Ich wollte herausfinden, ob und wie sich diese Unterschiede auf die Benennung der Fabelwesen auswirken. Hauptsächlich zeigte mein Experiment, dass polnischsprachige Teilnehmer weniger intensiv in ihren Phonemen differenzierten ...

Morpheme expansion and linguistic mistakes of the Russian government

The ongoing invasion of Ukraine has significantly affected Russian language and its use in public space — certain actions by the Russian government have resulted in an unintentional morpheme expansion, essentially forcing the performative aspect t...

This talk is based on the essay written for the "Sprache und Krieg" Seminar and deals with the transformation of the role of language in Russian propaganda and specifically the unintended morpheme expansion caused by the hastily actions of the Russian government and police.
I have examined the previous examples of Russian propaganda as well as the documents prepared for the expected successful outcome of the invasion in Ukraine and compared the use of language there with other developments that occurred post-invasion.

This talk does not present a deep linguistic analysis but might be of use to anyone interested in the linguistic developments related to war, protests and propaganda, presenting data on the side of the current situation some might not be aware of due to the language barrier.

Implementing a Chatbot - Rule Based vs. Machine Learning

Tools for implementing chatbots can generally be divided into two categories - rule based and machine learning systems. This presentation is about a bachelor thesis project that compared the implementation process of a customer service chatbot usi...

Customer service chatbots can be implemented using either a rule-based system or one based on machine learning. My bachelor thesis compares the system developed by Kauz Linguistic Technologies (an example of a rule-based system) with Dialogflow by Google (an example of a machine learning-based system) from both a developer perspective and the perspective of the end user. Each system was used to implement a German-language chatbot for the use case of a consultation for a bank account. The platforms were then evaluated using a systematic framework. In a pilot user study, 20 participants (mean age of 23, 45% female and 55% male) evaluated the resulting chatbots in terms of usability, usefulness and satisfaction. With regard to the implementation process, the Kauz system was found to have more built-in functionality as well as data structures that make it easy to make changes to the finished chatbot while Dialogflow was easier to troubleshoot and made it simpler to extract and remember information from user queries. In the user study, while ratings were generally positive for both chatbots, participants showed a preference for the chatbot developed with the Kauz system.

Innovation or Retention? On the [ɵ] > [ɐ] sound change in Yue dialects

Dialect variation offers us huge insights in the history of a language. Through looking at the geographical distribution of dialect variants and comparing inter-dialect group variation, we have gathered evidence to propose a previously unattested ...

Abstract:
Within Yue and Guinan Pinghua dialects spoken in Guangdong and Guangxi provinces in southern China, there is a synchronic correspondence between the vowel [ɵ] and [ɐ]. For the dialects which have the rhymes [-ɵn] and [-ɵt], they form a minimal pair with -ɐn and -ɐt (e.g. [tshɵt5] ‘out’ vs. [tshɐt5] ‘seven’), meaning these dialects have as extra phoneme /ɵ/.

Our study is based on the data available in the Yubao online database (Zhongguo Yuyan Ziyuan Baohu Yanjiu Zhongxin 2022), with 78 Yue and Guinan Pinghua. We specifically looked at words which are pronounced as [-ɵn] and [-ɵt] in the Guangzhou dialect; this class of words consists of 15 characters in the database.

By looking at the geographical distribution of the presence of the ɵ : ɐ contrast, we found that the majority of the Yue dialects do not show such contrast. An interesting areal pattern emerged here is that /-ɵn/ and /-ɵt/ can only be found near Guangzhou, Zhongshan, as well as Hong Kong and Macau, but they are surrounded by dialects which do not show the contrast. In this paper, working with the assumption that the areal pattern we see is the result of a sound change, we would like to explore the f...

The role of communication in the development of sign language lexicons

The most learnable language would be one with an identical form for every word. But the interactive nature of languages means they are under constant pressure to use distinct forms for distinct meanings, so as to allow for successful communication...

To represent an object iconically with your hands, gesturers (people who do not have a sign language) overwhelmingly mime the *use* of the object, shaping their hands as though holding and manipulating it (‘handling’ strategy). In contrast, sign language users are much more likely to shape their hands as though their hand *is* the object (‘instrument’ strategy). I hypothesised that this difference is due to the handling strategy having a greater risk of homophony (one form for multiple meanings), leading signers to favour the instrument strategy in order to communicate effectively.

To test this, 491 participants were shown signs for a range of meanings, using one of the two strategies (handling or instrument), and asked what they thought the meaning was. This talk describes the results of this experiment, in particular comparing how semantically varied interpretations were, for handling and instrument signs. A narrow range of interpretations would suggest high communicative efficacy for that sign, and reduced risk of homophony in a sign language lexicon using that sign.

While the primary finding was no overall difference between strategies, some words individually had a si...

An experimental approach to principle C in German

Standard binding theory (Chomsky 1981) proposes that coreference and binding are regulated by binding principles A, B, and C. Additionally, it was generally accepted until a few years ago that coreference and binding are regulated by c-command (Reinhart 1976, 1983), a syntactic principle.

Definition 1: Principle C: All R-expressions must be free.

Definition 2: C-command: Node A c(onstituent)-commands node B iff the branching node α₁ most immediately dominating A either dominates B or is immediately dominated by a node α₂ which dominates B, and α₂ is of the same category type as α₁.

Principle C and c-command taken together predict why coreference between the pronoun preceding the R-expression Lucy is possible in (1a), but not in (1b).

(1) a. Her₁ sister found Lucy₁.

b. *She₁ found Lucy₁.

However, there are many counterexamples showing that non-syntactic factors can influence coreference judgments (Bolinger 1977) or even obviate principle C, such as etiquette requirements in (2).

(2) (Schlenker 2005: 399, ex. 37)

[The King of T...

Hate Speech Detection on Code-Switched Data

The following presentation will focus on adapting a pre-trained BERT model to a code-switched dataset to be able to create a hate speech detection model handling code-switching.

With half of the human population having a social media account, it has never been easier to express your thoughts and opinions in front of a wider audience. As a consequence, verbal aggressions such as hate speech have increased dramatically and are not only harming the mental health of individuals but also damaging the relationship between different ethnical groups due to offensive and abusive content. Therefore it became important to lawmakers and social media platforms to filter this kind of behaviour online. For instance, each social media platform has its own policies to remove hate speech based on their definition. While Facebook defines hate speech as violent or dehumanising speech or calls for exclusion or segregation of people based on protected characteristics such as race, gender, etc., Twitter sums up everything which promotes violence or threats against other people. In the past, this kind of content has been removed manually, but with 500 million tweets and 4.3 billion Facebook messages everyday, manual hate speech detection has become impractical and ineffective. Consequently, an automated approach for hate speech detection is required. Therefore several machine ...

Go small or go big, there is no in-between

Code-switching is a language phenomenon where multilingual speakers alternate their use of languages in one utterance. While extensive studies have been conducted on Spanish-English speaker communities (Anderson & Toribio, 2007; Badiola et al., 2018; Poplack, 1980), the Mandarin-English bilingual population, despite its large size, remains understudied. As some English grammatical elements do not have direct Mandarin equivalents, Mandarin and English make an interesting pair for code-switching grammar studies. This study investigates the grammar preferences of Mandarin-English speakers in code-switching. In the present study, two specific grammatical elements were selected for comparison: plural marker (-s, -es) and determiners (a, the, etc.). Twenty pairs of code-switched sentences contrasting in the inclusion or exclusion of these grammatical elements were created for a comparison task. 100 Mandarin-English bilingual speakers completed a survey in which they described their language background, provided judgments for the comparison task, and indicated their attitude toward code-switching behaviors. While evidence compiled from the result supports the Matrix Language-Frame mode...

The binding of German anaphors in a Minimalist Framework

The classic Conditions A and B of Binding theory do not sit well with the current Minimalist approach to grammar, which aspires to eliminate all phenomena-specific concepts and replace them with analyses fully reliant on a small number of generally applicable, independently motivated grammatical principles.
In his dissertation, Glyn Hicks (2009) develops a Minimalist account of anaphoric and pronominal binding, attributing the effects to basic syntactic principles and thus is able to eliminate the binding conditions from the framework. He extends this account to various Germanic languages and their split anaphoric systems, which consist of SE as well as SELF anaphors (e.g. Dutch 'zich' and 'zichzelf').
German, which is not addressed by Hicks, also has two anaphoric elements, 'sich' and 'sich selbst', which differ greatly in their behaviour from every anaphoric inventory analysed in the dissertation (see for example Hendriks et al. 2014). The given system cannot easily account for these data.
I propose that by treating German 'sich' essentially not like a SE, but like a SELF anaphor with regards to syntax, the data can largely be reconciled with Hicks' framework, maintaining ...

When iconicity diverges

Sign languages, the manual-visual languages of deaf communities, exhibit a high degree of iconicity in their vocabularies due to the affordances of the modality for creating visually motivated form-meaning mappings (1). The potential for iconicity...

Past studies have explored how iconicity influences the lexicon in the visual modality, typically paying limited attention to signs’ formational components and their cross-linguistic differences (4–9). Other studies have conducted lexical comparisons across sign languages to establish cross-linguistic similarity, typically discarding iconic signs from the analysis (10–14). Yet, a deeper understanding of the factors that shape a manual lexicon must include how these two forces interact in sign language lexica: iconicity driving cross-linguistic similarity and language-specific conventions driving differences. In the present study, we investigate this interaction, by analysing types of iconicity (i.e., strategies to realise form-meaning mappings) and degree of form overlap in BSL and DGS.

We selected up to 10 concepts (N = 234) (concrete and abstract) elicited from deaf signers in a semantic fluency task in DGS and BSL for analysis. We classified all signs by iconic strategy (enactment, acting on, representing, personification, entity, tracing, moulding), are collecting iconicity and concreteness ratings, and establishing edit-distance, based on the formational components, for ...

Open session

The open session allows participants to meet some of the sponsors of the joint 71st StuTS + 31st TaCoS as well as interact with each other.

Posters:

See links for more information on the posters.

Aykut Guler: Internet Memes as Argument Prompts - Roses are red, violets are blue, internet memes are arguments too

Stefania Pikus: Phonetic Convergence in German non-native speech

Meeting Junge Sprachwissenschaft e.V.

Internet Memes as Argument Prompts

What is an argument? When exactly does someone argue? What is the role of the audience while the argument is formulated? In this paper, I have discussed two perspectives of argumentation: argument as a cognitive category and argument as a verbal practice. The distinction is an important one to make, because the former perspective implies that arguments invite a guided interpretation of a message whereas the latter perspective sees the argument as the message itself. As such, argumentation as a cognitive category enables multimodal media to be analyzed alongside traditional verbal propositions. I lend support to this cognitive stance, which I show via an argumentative analysis of two Internet Memes, using the framework proposed by van den Hoven & Schilperoord (2017) in their analysis of editorial cartoons. In doing so, this paper demonstrates that (1) Internet Memes can also be deliberately argumentative, and it is possible to analyze them systematically; (2) The model of van den Hoven & Schilperoord (2017) can have further applications in argumentation studies.

Bibliography
van den Hoven, P., & Schilperoord, J. (2017). Chapter 5. Perspective by incongruity: Visual argumenta...

Meeting STReNeL

Our goal is to achieve better and more reliable scientific advances and generalizations, by thinking about solutions to the urging replication crisis. By means of educating fellow young researchers, we aim to form a collaborative network that tack...

Our goal is to achieve better and more reliable scientific advances and generalizations, by thinking about solutions to the urging replication crisis. By means of educating fellow young researchers, we aim to form a collaborative network that tackles issues that are related to the replication crisis.

LangSci Press

Meeting ILLSA

Underlined

Dutch Language Institute

Pub Crawl

Breakfast @ accommodation

Spraakherkenning, wa is da? - Bias in Flemish Speech Recognition

Previous research has touched upon bias in automatic speech recognition (ASR) but has mostly focused on the English language. However, as these systems are used around the globe, finding biases in different languages is of high importance. The current study seeks to find biases based on age, region, and sex in a Dutch state-of-the-art ASR model. The Dutch data from the Spoken Dutch Corpus will be used to train the hybrid DNN-HMM model and it will be evaluated using the Flemish data from the JASMIN-CGN corpus. In addition, I will provide an in-depth phoneme analysis. The findings will contribute to a better understanding of bias, and subsequently inclusivity, in ASR.

Not gonna lie - An investigation into constructional change

The work-in-progress presented in this talk is based on the ideas of Diachronic Construction Grammar and aims to connect the basic principles of constructional change, more precisely the idea of chunking, outlined in Schmid (2020) with what is known about the composition of prosodic units. As an example, this study will use the phrase not gonna lie and show how the compositional phrase "AGENT BE not going to lie to PATENT" (e.g., "I’m not going to lie to you") became the internet initialism "ngl", using data from the "Corpus of Contemporary American English" (COCA).
The study will first use traditional Construction Grammar analysis for the individual stages of the construction (based on the Goldberg (2006) definition of constructions), to show how the phrase becomes less compositional over time and how the meaning and use of the phrase changes. Furthermore, since these types of analyses tend to emphasize the elements that become fixed, the talk will put particular focus on the elements that are omitted during the development of the construction. As a last point, the talk will show how the parallel reduction theory (Bybee et al. 1994) and the prosodic structure of intonation phr...

“Don’t speak Latin in front of the books!”: An exploration into the linguistic influences on the creation of magic systems in fiction

There are numerous examples of magic systems in fiction, many of which involve spellcasting by an agent. Spellcasting often involves incantations, which are verbal components that a spell or magical effect requires in order to be realised or cast (sometimes supplemented by somatic components). This project examines the influences on and and choices made for the linguistic representation of incantations, in a small corpus that aims to be somewhat representative of the fantasy genre. This is also achieved through choosing a range of media, such as films, books, and games. Furthermore, the corpus includes media that is specifically aimed at adults, among which are Shakespeare’s Macbeth and Joss Whedon’s Buffy the Vampire Slayer, and media that is specifically aimed at children, among which are Hayao Miyazaki’s Spirited Away, and Ursula K Le Guin’s Tales from Earthsea. In addition to examining incantations with an English basis, I focus somewhat on the significance of constructed languages (conlangs) used in magic systems, and the factors that influence the choice to use a conlang to represent the verbal counterpart of a magic system.

My analysis of the corpus reveals strong infl...

Information structure in speech synthesis

The term ‘information structure’ refers to a way of forming a message in a way that makes it most easily understandable for the hearer against the background of a certain context (Féry & Krifka 2008). Two examples for information-structural status...

To test this assumption, I will investigate whether the naturalness of a German synthetic voice can profit from an intonation that matches the focus structure of the sentence. Based on Domínguez et al. (2018) who have shown that the mean opinion score (MOS) for a synthetic voice improves when the prosody is automatically or manually modified in accordance with thematicity, I hypothesize that the MOS for a German synthetic voice will also improve when the prosody is modified to match the focus structure.

To that aim, I am planning to synthesize two German voices - one with and one without a focus-matching prosody - and compare their MOS’s. To create these voices, I will either train an existing statistical-parametric-speech-synthesis (SPSS) based TTS system once with focus/non-focus labels and once without focus/non-focus labels or I will adopt the strategy of Domínguez et al. (2018) who used an open-source TTS software and enriched the default synthesized speech with prosodic information according to the thematicity of the corresponding sentence parts.

A result that confirms my hypothesis would suggest that information-structural categories should be included in the creati...

Semantic Differentiation of Old Irish Terms for Bodies of Water

This study investigates the semantics of Old Irish terms in the lexical field 'body of water' based on a corpus of electronically available texts. The semantic investigation combines corpus linguistic approaches with close readings of relevant tex...

Words do not create meaning in isolation, but have complex semantic relationships with other words; additionally, languages avoid having multiple words with the exact same meaning and context of use (Kay and Allan 2015). Semantic investigations, therefore, benefit from the analysis of entire lexical fields in order to identify and explain semantic differences. Historical semantic investigations often rely on corpus linguistic methods, as the surviving material constitutes a corpus. The limitations of the surviving Old Irish corpus require a combination of quantitative and qualitative methods (Lash, Qiu, Stifter 2020). The semantics of the Old Irish lexical field ‘water expanse’ have previously been investigated by Liam Mac Mathúna, who based his investigation on dictionary entries and citations (1987).
Corpus linguistics and electronic availability of Old Irish material have improved since that study, so that a larger corpus can now be examined. This warrants a re-examination of the lexical field. The present study, therefore, investigates the semantics of terms in the lexical field ‘body of water’ in order to differentiate their semantics. This provides insight into the semant...

#GermanisTikTok

„wer benutzt ÜBERHAUPT komma“ – Diese Frage einer TikTok-Nutzerin in der Kommentarsektion unter einem Video, das sich inhaltlich mit dem Thema Kommasetzung beschäftigt, hat mich Anfang des Jahres dazu bewogen, ein aufklärendes Video über korrekte ...

„wer benutzt ÜBERHAUPT komma“ – Diese Frage einer TikTok-Nutzerin in der Kommentarsektion unter einem Video, das sich inhaltlich mit dem Thema Kommasetzung beschäftigt, hat mich Anfang des Jahres dazu bewogen, ein aufklärendes Video über korrekte Kommasetzung auf ebendieser beliebten Social Media Plattform hochzuladen. Knapp zwei Monate später habe ich mit meinem Account @clebriclips über 50.000 Follower und mehrere hunderttausend Aufrufe erreicht.
Mein Name ist Clemens Brill, ich bin Student der Theaterwissenschaft und der Germanistik an der Johannes Gutenberg-Universität in Mainz und stehe kurz vor meinem Bachelorabschluss im 6. Fachsemester. Seit Januar 2022 bin ich auf GermanisTikTok aktiv und produziere kurze Videos mit maximal drei Minuten Länge, in denen ich sprach- und literaturwissenschaftliche Themen erkläre. Eindeutiger Publikumsliebling: das Thema Kommasetzung.
Wie genau funktioniert eigentlich Kommasetzung? Was ist eine Subordination? Steht vor jedem ‚dass‘ ein Komma? – Fragen, die scheinbar viele Nutzer*innen der App beschäftigen, die durch eine unzureichende Kommadidaktik in der Schule und weitestgehend der Universität Wissenslücken in diesem Bereich zu haben ...

BuFaTa

"We were able to communicate after all.”

As the world becomes increasingly interconnected and mobility increases, more and more encounters take place in which people interact despite the absence of a shared named language in their linguistic repertoires. Such communicative encounters may...

In this paper, I therefore present a close interactional analysis of an interaction without shared named languages, examining the communication strategies (Selinker, 1972; Dörnyei & Scott, 1997) and resources (Mondada 2019; Kidwell 2013) employed. The analysis is based on an audiovisual recording of a simulated restaurant visit in which four South Tyrolean secondary school students interacted with a Russian waitress and a tourist from Morocco seated at the same table. While all four students disposed of several named languages and varieties in their linguistic repertoires, none of them spoke Russian or (Moroccan) Arabic. In this paper, I investigate the communication strategies and semiotic resources which the interactants employed during this task, such as the use of deictic gestures and particles to establish a common focus and the use of objects and representational gestures in combination with speech to make verbal resources available to the co-interactants. In total, I show that interactants cooperatively establish common ground through the reflective, constantly adapting management of all available resources.

References:
Dörnyei, Z., & Scott, M. L. (1997). Communicati...

Language and attachment in mother-child interactions: Relations between related talk and attachment

In my master’s thesis, which combines psychological and linguistic theories, I analyse the relationship between maternal attachment style and the mothers' use of language to talk about the actions and speech of their children (related talk). My hy...

In the 70ies of the last century, Mary Ainsworth and Mary Main described four different patterns in child-parent attachment. These four attachment styles have also been identified in adults. Broadly speaking, insecurely attached individuals have difficulties building and maintaining close relationships while securely attached adults have less problems. In my data, which consists of 40 mother-child dyads (age of child: 5 years) solving a task together, the attachment type of the mothers has been identified. My main research question is: To which degree do mothers with different attachment styles use their language to talk about the actions and speech of their children (vs. about other things)? For this purpose I have developed and applied a developmentally and situationally appropriate concept of contingent talk/verbal responsiveness which I have termed "related talk". This allows me to categorize all utterances by the mother as either related to the current action or speech of the child or not. Preliminary results show that securely attached mothers use referential talk more frequently than insecurely attached mothers.

Existential Generics and Information Structure

Generic sentences are used to convey exception-tolerant generalizations over a kind or, roughly, relevant members thereof (Carlson & Pelletier 1995; Cohen 2016). In English they often take the form of a bare plural (BP). The conditions under which...

Evidence in Support of a Cognitive Bias for Cross-Category Harmony between the Verb Phrase and the Adpositional Phrase in the Absence of Surface-Level Patterns

The vast majority of languages for which we have the data display syntactic harmony between the verb phrase and the adpositional phrase. Previous work suggests this is due to the cross-linguistic tendency for adpositions to derive from verbs. Rese...

In this talk I describe three experiments designed to establish whether there is a cognitive bias for harmony between the adpositional phrase and the verb phrase, when the possibility of using surface-level patterns is removed. These used a combination of artificial language learning and silent gesture perception methods. Experiment 1 investigated the possibility of a baseline preference for adpositional phrase order in English-speaking participants, because it is the extent of their preference for particular adpositional orders which is manipulated in experiments 2 and 3. Results showed no evidence for a baseline preference for prepositions or postpositions. Experiment 2 showed that the experimental methods employed are sufficient to demonstrate a harmony effect when the elements in the priming and testing stimuli are the same. This replicates results of previous experiments which utilised different methods. Experiment 3 revealed a preference for harmonic patterns is also present when there are no repeated elements in the priming and testing stimuli, to a lesser extent than in experiment 2. This indicates that there is a bias for harmony between the verb phrase and the adpositi...

Multilingualism in Crisis Communication

This on-going project investigates the availability, accessibility, and acceptability of multilingual and simple COVID-19 information for individuals in the Netherlands who have a low mastery of written Dutch either due to another mother tongue or...

With most of the COVID-19 communication being done in Dutch or using English as a linguafranca, minority language speakers, immigrants, and people with lower (or no) literacy, face the challenge of linguistic barriers that make them more vulnerable to the virus and its consequences(OECE, 2020). 
This project aims to create a clear image of how the role of multilingualism plays in the Netherlands during COVID-19 pandemic. This project taps into and is rooted within different linguistic traditions that it uniquely combines and integrates within the context of a timely and relevant topic: Crisis sociolinguistics, linguistic diversity, and linguistic landscape. Observational data will be used to provides an overview of the current multilingual information available concerning COVID-19. Furthermore, to have a comprehensive image of the role of multilingualism in the Netherlands during the pandemic, three groups of participants are recruited nationwide: non-Dutch speakers, low-literate L1 Dutch speakers, and a control group consisting of literate L1 Dutch speakers. Questionnaires and interviews are performed to compare the availability and immediacy of public information to people of...

Investigating the semantics of tense-aspect markers in Jula

Our talk aims at presenting a research project that we did on the tenses and aspects markers of Jula (an understudied Mande language).
We will show how to make a distinction between tenses and aspects on the theoretical level, and how to investig...

Does contrast affect the use of the personal pronoun er and the pronominal demonstratives der and dieser in German?

Two of the most frequently used German demonstratives are der and dieser, which can be used adnominally (like determiners) and pronominally (like pronouns). Although there is some recent work on the pronominal use (Patil et al. 2020, Fuchs & Schumacher 2020), it remains unclear which linguistic factors influence a speaker’s preference to use der or dieser. One of these factors might be contrast, which has been discussed theoretically but not empirically tested. In this presentation, I report the results of two forced-choice experiments in which I empirically investigated the question whether the linguistic factor contrast affects people's choice to use the German personal pronoun er or the pronominally used demonstratives der and dieser. Bisle-Müller (1991) and Ahrenholz (2007) assume that only dieser and not der is able to express a contrast between several referents and therefore should be preferred in those contexts. Contrary to that, Bosch & Hinterwimmer (2016) argue that der is able to express contrast, which might be a difference between der and the personal pronoun er. The statistical analysis of the results of the two experiments shows that der and dieser seem to be equa...

Die Auswirkungen des Krieges auf die Sprache: Lehnwörter im Englischen

Im 20. Jahrhundert gab es mehr als 200 Kriege auf der Welt und fast 80 Millionen Menschen sind in Kriegen gestorben. Als Ausdruck des sozialen Wandels brachten Kriege nicht nur Leid und Katastrophen über die Menschheit, sondern katalysierten auch Revolutionen in anderen sozialen Phänomenen, wofür die Entwicklung der Sprache ein typisches Beispiel ist. In diesem Aufsatz wird die Beziehung zwischen Kriegen und lexikalischen Entlehnungen im Englischen aus dem Französischen diskutiert. Insbesondere wird das Phänomen der lexikalischen Entlehnungen zwischen dem Französischen und dem Englischen im Hinblick auf die Kriegsgeschichte zwischen diesen beiden Ländern analysiert; darüber hinaus werden in diesem Aufsatz praktische Belege präsentiert. Die eingehende Analyse zeigt, dass das Phänomen der englischen Entlehnung französischer Wörter in fast allen Bereichen des täglichen Lebens auftritt und dass die meisten dieser Entlehnungen auf Kriege über Jahrhunderte hinweg zurückzuführen sind. Es zeigt sich auch, dass die lexikalischen Entlehnungen zwischen dem Französischen und dem Englischen zweiseitig sind, während das Französische auch durch Lehnwörter aus dem Lateinischen durch Konflikte u...

Metapragmatische Marker als Konfiguratoren mentaler Räume

Slides available on request. Please get in touch: oliver.spiess@unibas.ch

Metasprache im Allgemeinen und Metadiskurs/Metapragmatik im Besonderen haben in den vergangenen Jahrzehnten als Forschungsgegenstand der Linguistik an Bedeutung gewonnen (Caffi 2016:799; Hyland 2017:16). Trotzdem, oder gerade deshalb, wird von vielen Forscher:innen mittlerweile konstatiert, dass keine Einigkeit darüber besteht, als was Metadiskurs/Metapragmatik genau zu definieren wäre (Ädel 2006:3; Hyland 2017:16; Techtmeier 2001:1449; Martínez Guillem 2009:731). Zur Charakterisierung von Metadiskurs/Metapragmatik wird oft die definitorische Bedingung des Bezugs der entsprechenden Form auf den "evolving discourse", "ongoing text" (beide: Ädel 2006:2), "current discourse" (Ädel 2006:27; Hübler & Bublitz 2007:4), "ongoing discourse" (Ädel 2006:28; Hübler & Bublitz 2007:6; Caffi 2016:801; Zhang et al. 2017:106) oder auch die "ongoing interaction" (Liu & Ran 2016:285) genannt. Dabei scheint gar nicht per se klar, was zum aktuellen Diskurs gehören soll, denn "die Kriterien, die dafür benannt werden – Einheit von Raum und Zeit, Konstanz der lnteraktionsbeteiligten, Funktion und Thema, institutionelle Begrenzung etc. sind alle nicht unproblematisch" (Techtmeier 2001:1452).
...

Argumentative Muster von jungen CoVid-Skeptiker*innen und -Leugner*innen in Österreich.

Seit über zwei Jahren prägt die CoVid-19-Pandemie weltweit gesellschaftliche, politische und mediale Diskurse und nimmt damit auch Einfluss auf unseren Sprachgebrauch. Neben der Entstehung bzw. Neukonnotation konkreter Lexeme („CoVid-19“, „Quarant...

In der präsentierten Arbeit wurden daher die argumentativen Strukturen unter jungen CoVid-Skeptiker*innen und –Leugner*innen in Österreich untersucht. Als Stichprobe wurden dazu sämtlich öffentlich zugänglichen Telegramgruppen der Bewegung „Studenten stehen auf“ zugezogen, die sich durch eine eindeutige Gruppenbenennung jeweils einer von 8 österreichischen Städten zuordnen. Untersucht wurden Nachrichten aus dem Zeitraum des 4. gesamtösterreichischen Lockdowns zwischen 22. November und 12. Dezember 2021, die mindestens einen argumentativen Schluss aufweisen, der eindeutig durch eine Prämisse und eine Konklusion gekennzeichnet ist. Das nach diesen Kriterien erstellte Korpus wurde mithilfe der „Extensive Markup Language“ (XML) strukturiert und anhand der von Toulmin (1996) und Schröter (2021) konzipierten linguistischen Argumentationsanalyse annotiert. Die Ergebnisse beschreiben die argumentative Makro- und Mikrostruktur der analysierten Nachrichten, wobei die Mikrostruktur und damit die verwendeten Prämissen und argumentativen Topoi im Vordergrund der Untersuchung stehen. Konkret werden die Fragen beantwortet, welche Argumentationstypen verwendet, welche Topoi dabei etabliert bzw....

Pragmatic Markers in Direct Messaging

A comparative study of pragmatic markers used in direct messaging and oral conversation.

In my bachelor’s thesis I analysed the use of pragmatic markers in direct messaging. While pragmatic markers are frequently used in oral, synchronous conversations, either face-to-face or on the phone, they appear less often in written, asynchronous language. Direct messaging in current times often happens quasi-synchronous and often contains oralized written language. Therefore, it is interesting to investigate the use of pragmatic markers in this context. A corpus of text messages was compiled for this study. Both ‘classic’ pragmatic markers, such as ‘well’ or ‘just’ were analysed, as well as some net-speak acronyms such as ‘lol’ or ‘lm(f)ao’, to see whether these acronyms can be used as pragmatic markers. The findings in the text message corpus were compared to Kate Beeching’s (2016) findings. The results of this study as will be presented in this talk.

Phonetic Convergence in German non-native speech

The analysis is based on the project entitled "Automated analysis of phonetic convergence in speech technology systems". The project of the Faculty of Modern Languages of the Adam Mickiewicz University in Poznań was supported by the grant Harmonia...

To ascertain whether or not there was an increased measurable convergence in non-native speech and how that was influenced by the interlocutor and the outer circumstances, non-native speakers were put in diverse dialogue scenarios. The test subjects were ten native speakers of Polish. They were asked to first read and afterwards repeat four utterances after a test moderator who was a native speaker of German. There were five tasks included in the experiment. A reading task, where the speakers had to read out loud the sentences including the target utterances, which was done separately for each participant, and four speaking tasks in pairs. The correctness of the sounds was determined by audiovisual analysis. There is an overall improvement tendency. Improvement is visible in the sounds /e:/, /x/, and /œ/, while there is a decline in the longer sound /ø:/. Since the sound /e:/ is in different places in the sentence, there is a different improvement ratio. The range of improvement in the sound /x/ was from 40-50% to 90-100%. In the case of “ö” pronounced as /ø:/ and /œ/ the variation of the sound might have caused problems. In this sound pair /ø:/ was pronounced as /œ/ by three p...

Closing Plenary

Fries Straatfestival

Brunch @ Lazy Lemon

Ruiterskwartier 135A
Leeuwarden