Thesis Topics for Master Students

Many speakers who use gender-neutral speech in German produce a glottal stop /ʃtʊdɛntʔɪn/. In this project you will compare the realization of gender-neutral speech for different speakers (those who find this distinction important and those who don’t). You will also investigate how the speakers deal with forms in which the stress pattern is different between the male and female forms (e.g., Professor vs. Professorin)

Prerequisites:

• knowledge of praat or another speech processing software
• knowledge of statistics

Language: 

German or English

In mother-child interaction, mothers utter a whole range of semantically empty one-word utterances (e.g., hm?). In this project you will analyse an annotated corpus of spoken Germna mother-child interaction to classify the pragmatic content (intent) of these one-word utterances and to relate their function to intonational form to develop a model of intonational meaning.

Prerequisites:

• knowledge of praat or another speech processing software
• knowledge of statistics
• knowledge of pragmatics

Literature:
Grice, M., Baumann, S., & Benzmüller, R. (2005). German Intonation in Autosegmental-Metrical Phonology. In J. Sun-Ah (Ed.), Prosodic Typology. The Phonology of Intonation and Phrasing (pp. 55-83). Oxford: Oxford University Press.

There are studies showing that participants perceive high-pitched syllables as stressed (this happens, for instance, in a rising intonation pattern such as “Peter?” where the low-pitched stressed syllable is less often detected correctly than in a declarative utterance with a falling f0-contour. In this project you will use rhythm-beats to probe participants’ stress perception in different intonation contours.

Prerequisites:

• knowledge of praat or another speech processing software
• knowledge of psycholinguistics
• knowledge of statistics

Literature:
Zahner, K., Kutscheid, S., & Braun, B. (2019). Alignment of f0 peak in different pitch accent types affects perception of metrical stress. Journal of Phonetics, 74, 75-95.

Previous research has shown that listeners can use prosodic information (duration, voice quality, intonation) to decide whether a question is intended as rhetorical or information-seeking question (all else being equal). In this project you will investigate how this prosodic information compares to information from other linguistic areas, e.g. syntactic structure, use of particles

Prerequisites:

• knowledge of psycholinguistics
• knowledge of statistics

Literature:
Braun, B., Dehé, N., Neitsch, J., Wochner, D., & Zahner, K. (2018). The prosody of rhetorical and information-seeking questions in German. Language and Speech 62(4), 779–807.
Kharaman, M., Xu, M., Eulitz, C. & Braun, B. (2019). The processing of prosodic cues to rhetorical question interpretation: Psycholinguistic and neurolinguistics evidence. Proceedings of Interspeech. Graz, Austria.

Previous research has shown that speakers mark their own bias preferentially via word order in German (Gibt es nicht einen Bus? vs. Gibt es keinen Bus?), while English speakers make more use of prosody (in particular the final rise). In this project you test the perception and identification of speaker bias and test the roles of prosody and syntax therein.

Prerequisites:

• knowledge of psycholinguistics
• knowledge of statistics

Literature:
Arnhold, A., Braun, B. & Romero, M. (2020). Aren’t prosody and syntax marking bias in questions? Language and Speech. Online first publication  https://doi.org/10.1177/0023830920914315

Rhetorical questions (RQs) may be signaled by lexical or syntactic cues and/or by prosody. Regarding the prosodic marking of RQs, previous research on German has shown that tonal targets are aligned later in rhetorical wh-questions than in neutral wh-questions. Swiss German is an interesting test case for the marking of RQs as tonal alignment seems to occur later in Swiss German than in Standard German. Moreover, Swiss German employs various lexical cues to convey pragmatic meaning. In this project you will test whether Swiss German speakers use syntactic/lexical cues and/or prosody to mark a question as rhetorical. If prosody is a cue, you will further study tonal alignment patterns for the disambiguation of the two illocution types in more detail.

Prerequisites:

• knowledge of intonation
• basic knowledge of syntax
• knowledge of praat or another speech processing software

Literature:
Braun, B., Dehé, N., Neitsch, J., Wochner, D., & Zahner, K. (2018). The prosody of rhetorical and information-seeking questions in German. Language and Speech 62(4), 779–807.
Leemann, A. (2012). Swiss German intonation patterns. Amsterdam / Philadelphia: John Benjamins Publishing Company.

Develop a Grammar Fragment for a language using LFG/XLE.

Computationally analyze linguistic strategies involved in framing and argumentation. Other possible topics are the automatic detection of hate speech or the content of political manifestos.

Develop a computational analysis for morphophonological phenomena in a language using Finite-State Morphology.

Work on a topic within Natural Language Unterstanding: the automatic analysis semantic content.

Develop small AI systems. These could be Chatbots or systems involving machine learning from texts for a given task like text generation or classification or clustering of texts/phenomena.

I am happy to supervise topics on the following: case, complex predicates, lexical semantics, grammar architecture, including the prosody-syntax-semantics/pragmatics interface. My area of specialization is South Asian languages, but I am happy to do other languages as well.

Understanding language change, particularly with respect to case or auxiliary formation. This can be done purely from a general linguistics perspective or be combined with computational approaches (corpus linguistics and/or visual analytics).

In this thesis, you will examine the instances of emotion and emotion-evoking language in Spanish political manifestos and speeches. The aim is to create word lists that will help us in the analysis of emotional language in Spanish texts. 

Required skills:

• knowledge of Spanish; no programming skills required

Previous research has shown that listeners can use prosodic information (duration, voice quality, intonation) to decide whether a question is intended as rhetorical or information-seeking question (all else being equal). In this project you will investigate how this prosodic information compares to information from other linguistic areas, e.g. syntactic structure, use of particles

Prerequisites:

• knowledge of psycholinguistics
• knowledge of statistics

Literature:
Braun, B., Dehé, N., Neitsch, J., Wochner, D., & Zahner, K. (2018). The prosody of rhetorical and information-seeking questions in German. Language and Speech 62(4), 779–807.
Kharaman, M., Xu, M., Eulitz, C. & Braun, B. (2019). The processing of prosodic cues to rhetorical question interpretation: Psycholinguistic and neurolinguistics evidence. Proceedings of Interspeech. Graz, Austria.

Literature on the intonation of Faroese is scarce, except for some short and anecdotal descriptions in Lockwood (1977) and Árnason (2011). Speech data (from a map task study carried out in 2019) are available for analysis.

Prerequisites:

- Knowledge of Praat, some experience with intonational annotation / analysis
- Interest in Faroese

Literature:

Árnason, Kristján. 2011. The Phonology of Icelandic and Faroese. Oxford University Press.
Lockwood, W. B. 1977. An Introduction to Modern Faroese (3rd printing). Tórshavn: Føroya Skúlabókagrunnur.

The objective of your thesis will be to explore important issues relating to ditransitive verbs in Insular Scandinavian. The project that your thesis will be related to focused on three main issues: (a) inversion of the two objects (DO-IO orders in active clauses and theme passives), (b) the morphosyntax of ditransitive verbs (different cases and DPs vs. PPs) and related syntactic issues, and (c) the scope possibilities for the internal arguments of ditransitive verbs.

Data have been elicited in a series of experimental studies at the University of Iceland, ready for analysis. If you are interested, we will contact colleagues in Iceland and discuss use of the data as well as a specific thesis topic.

Ich betreue Abschlussarbeiten im Bereich Semantik, Pragmatik und Sprachgeschichte. Sie können theoretische Arbeiten, Literaturvergleichende Arbeiten oder empirische Studien anstreben. Hier ist eine Auswahl an exemplarischen Themen. Für Ihre eigenen Vorschläge bin ich immer offen.

Most topics can also be researched for English, and in English.
See me in my office hours talk about your ideas and interests.

Mit perspektivierenden Ausdrücken wird die Meinung eines Sprechers wiedergegeben. Es gibt viele Formen der Perspektivierung: geschmacklich (gut, lecker, ekelhaft), emotional (leider, gottlob), epistemisch (wohl, vielleicht).

Da viele davon noch nicht genauer beschrieben wurden, können hier viele Einzelfallstudien durchgeführt werden..

Ein weiterer Aspekt perspektivierender Ausdrücke ist ihre Funktion in Medientexten. Welche Art von Mitteilung wird perspektiviert? Welche Perspektiven werden vermittelt?

• Datenbezogen: in welchen Texten und Medien werden rhetorische Fragen vermehrt verwendet?
• Form: Wie werden rhetorische Fragen im Deutschen markiert? (Negativ-Polare Elemente, Negation, Partikeln, Adverbien usw.)
• Funktion: Kann jede beliebige Frage auch als rhetorische Frage verwendet werden? Wie sehen adäquate Verwendungskontexte für rhetorische Fragen aus?

• Pragmatik von negativ-polaren Elementen
• Einordnung weiterer Negationsformen des Deutschen in die Theorie: Niemals, nicht einmal, nicht ein einzige(s) … und ihre Verwendung in Zeitungstexten
• Verwendung und Funktion von von wegen!

In den süddeutschen Dialekten sind die morphologischen Formen des Präteritums fast völlig verschwunden; sie werden durch analytische Formen im Perfekt ersetzt.

• welche bedeutung hat in diesem System das Doppelperfekt (Peter hat das Buch gelesen gehabt)?
• Überprüfung von Quellen des 16. Jahrhunderts auf die Verwendung von Perfekt / Präteritum
• Überprüfung von Quellen aus dem Bereich der Hanse auf die Verwendung des Perfekts / Präteritums im 15. und 16. Jahrhunderts
• eventuell für Romanisten: Quellen aus dem Norditalienischen Raum aus dem 14. Jahrhundert auf die Verwendung von Vergangenheitsformen hin überprüfen.

Erarbeitung von fallacies (= Scheinargumenten) anhand von Lehrbüchern, und eine empirische Studie zu der Frage: Welche fallacies kommen bei (online-)Debatten am häufigsten vor? (Zur Eingrenzung empfiehlt sich ein Fokus auf Debatten zu einem bestimmten Thema, z.B. nur über Migration, nur über Corona, …)

Wie wird in Texten die Perspektive des Autors indirekt spachlich vermittelt? Wie setzen Journalisten subjektive Prädikate ein, um einen gemeinsamen Glaubenshintergrund zu suggerieren?

Die Natur von A+N-Komposita im Russischen: Zur Semantik von A + N – Phrasen im Russischen. Im Deutschen bedeutet rote Socke dasselbe wie „ist rot und ist eine Socke“. Im Russischen scheint es A+N-Verbindungen zu geben, in denen das Adjektiv nicht einfach zur N-Bedeutung dazukommt. Was ist die Semanitk dieser A+N-Verbindungen? Gibt es solche Beispiele auch in anderen Sprachen (z.B. „blauer Brief“)? Wo ist die Grenze zwischen kompositionaler und konventioneller Bedeutung?

Befehle und auch manche Fragen können face-threatening acts sein. Oft wird behauptet, dass Sprecher auf indirekte Sprechakte ausweichen, um den FTA abzumildern. Ist diese Strategie empirische belegbar? Ist sie kulturspezifisch?

In this thesis, you will be conducting an EEG study using a component of the event-related activity, called MMN, to generalize MMN effects demonstrating the underspecification of phonological features in the mental lexicon.
This topic is conditional to the re-opening of the EEG lab.
Required skills: Experimental linguistics and neurolinguistics, ideally experience with EEG measurements and data analyses
Language of the thesis: English or German

In this thesis, you will be conducting an EEG study using a component of the event-related activity, called MMN, to investigate the pre-attentive recognition of the language mode of bilingual speakers while producing CV syllables in the L1 or the L2 as well as the nativeness of their productions.
This topic is conditional to the re-opening of the EEG lab.
Required skills: Experimental linguistics and neurolinguistics, ideally experience with EEG measurements and data analyses
Language of the thesis: English or German

You can come to me with ideas about projects on neurolinguistics topics. Language of the thesis: English or German

In this thesis, you will use regression methods to predict (extra)linguistic information (e.g., behavioral norms like concreteness, valency, arousal) using distributional semantic representations. The main focus of this project is to identify the type of linguistic and non-linguistic information available in distributional representations and understand how to access this type of information.
Required Skills: Understanding of distributional semantics; Understanding and running correlation and regression studies; Good knowledge of Python or R.
Language of the thesis: English only

In this thesis, you will use existing multimodal distributional representations (e.g., visual vectors) in combination with purely linguistic vector representations to understand how information coming from multiple modalities can be combined to solve traditional tasks from lexical and distributional semantics.
Required Skills: Understanding of distributional semantics; Understanding and running correlation and regression studies; Good knowledge of Python.
Language of the thesis: English only

In this thesis, you will build visual vectors using complex neural networks and test such representations on traditional tasks from lexical and distributional semantics. The main question in this project is the following: does the quality and type of pictures used to build such vectors significantly affect the final representation?
Required Skills: Understanding of distributional semantics; Understanding and running correlation and regression studies; Very good knowledge of Python; good knowledge of neural networks.
Language of the thesis: English only

You can come to me with ideas about projects that use corpus linguistic methods and distributional semantics to solve specific tasks.
Language of the thesis: English only

The aim of the project is identify what language policies are in place in countries with large multilingual populations in Europe and beyond.

This aim of this project is to identify through a survey the needs of the citizens of Konstanz in terms of the language policies they would like to be implemented in the future in Konstanz.
"Language policies in multilingual cities-2 and "Language policies in Konstanz as an international city – the view of the Konstanz citizens" are closely linked together.

The aim of this project is to find out through a survey with parents their perceptions about how social distancing has affected their children’s language development.

As part of the Ringvorlesung students are developing material for families and professionals. This project will evaluate the material through questionnaires and interviews with parents, professionals, people working in local authorities

The aim of the project is to study the effect of lexical stress on cognate production. While cognates are produced faster and more accurately than non-cognates (cognate facilitation effect (CFE)), it is unclear if and how lexical stress affects the CFE. A production experiment will be conducted to fill this research gap by testing German-English bilinguals on a picture naming task in both German and English. The pictures will depict non-cognates, cognates with stress overlap, and cognates with stress mismatch.

The aim of this project is to study how children produce and comprehend pronouns. Pronouns as referential expressions are crucial in everyday communication. The target like use of pronouns is considered to be a demanding task that costs a long developmental progress. It is little known about the strategies monolingual children are applying for producing and interpreting pronouns and even fewer studies investigate bilingual children strategies. In this project we address the question of pronoun use and interpretation by applying various tasks, e.g., story retelling, pronoun judgment task.

The aim of the project is to compare the comprehension of irony in monolingual (German) and bilingual children with Italian as a heritage language. Several tests will be used to assess the participants’ ability to understand irony, general cognitive abilities, and Theory of Mind (ToM). ToM is the ability to make inferences about other people’s beliefs, intentions, and states of mind. Your two objectives will be 1) to find out potential correlations between different tasks (within participants), and 2) to investigate two different dimensions of language acquisition (monolingual vs bilingual).

[Knowledge of Italian is not necessary; if present, the project can also include the heritage language.]

How do adults who learn German as a second language perform in Irony Comprehension tasks? Are there differences between groups with a different L1 (e.g., Italian vs Japanese)? In this project, you will be able to compare how Italian learners of German and another group of late learners perform in an Irony Comprehension task. The L1 of the second group of late learners can be chosen based on your linguistic experience.

The aim of the project is to test two groups of monolingual speakers, speakers of German and speakers of another language. The linguistic phenomenon under investigation is rhetorical questions, which involve different linguistic cues in different languages (syntax, lexicon, prosody). There will be three main experiments (perception, comprehension, production) and some additional tasks. Your objective will be to find out which cues trigger rhetorical interpretation in different languages.

The aim of the project is to compare heritage speakers of Italian and monolingual speakers of German on rhetorical questions which involve different linguistic cues in different languages (syntax, lexicon, prosody). There will be three main experiments (perception, comprehension, production) and some additional tasks. Your objective will be 1) to find out which cues trigger rhetorical interpretation in German, and 2) to investigate two different dimensions of language acquisition (monolingual vs bilingual).

[Knowledge of Italian is not necessary; if present, the project can also include the heritage language.]

Prof. Dr. Tamara Rathcke (with Prof.. Dr Theo Marinis)

The role of media in interpersonal accommodation and sound change has been controversially debated, with compelling evidence yet to be provided. In this project, the role of media engagement will be examined and compared between L1 and L2 speakers of English. This work will inform both sociolinguistic theory and second language acquisition models.

Prerequisites:

• Understanding of the posits of accommodation theory
•  Readiness to learn new technical skills
•  Basic knowledge of statistical inference

Literature:
J Stuart-Smith, G Pryce, C Timmins, B Gunter (2013). Television can also be a factor in language change: Evidence from an urban dialect. Language, 501-536.

Speed dating is one of the contexts that allows us to study verbal and non-verbal accommodation in highly relevant contexts. The time that speed-daters spend in each other’s company is very short, and what they say is often less important than how they say it. This project will study if and how communicative accommodation can explain and predict interpersonal attraction during speed dating. The data for this project was recorded at the Centre for General Linguistics in Berlin in mid-October, and is available in German.

Prerequisites:

• Good command of Praat or another speech processing software
• Basic knowledge of statistical inference

Literature:

Giles, H., & Ogay, T. (2007). Communication Accommodation Theory. In B. B. Whaley & W. Samter (Eds.), Explaining communication: Contemporary theories and exemplars (p. 293–310). Lawrence Erlbaum Associates Publishers.

If music/singing is your thing, you may want to consider a dissertation topic on shared structures of music and language. Both are uniquely human abilities, and their shared cognitive underpinnings have been controversially debated for quite some time. You can take up different stances on the issue at hand, but a potential project involves the study of the so-called “speech-to-song illusion”, an illusory perception of singing in speech. The effect has been documented in many intonation languages but is limited (if at all present) in tonal languages. There is also quite large, to date poorly understood individual variation in the susceptibility to the effect. You can run this experiment with your own data, or use existing recordings.

Prerequisites:

• Rock-solid knowledge of acoustic-phonetic concepts
• Understanding of key issues in speech and language processing
• Basic knowledge of statistical inference

Literature:

Falk, S., Rathcke, T. and Dalla Bella, S. (2014). When Speech Sounds Like Music. Journal of Experimental Psychology: Human Perception and Performance [Online] 40:1491-1506.

We can easily say who we perceive as charismatic, but it is more difficult for us to say why. Some previous research has shown that for a person to radiate charisma, the content of their message is not as relevant as the way the message is delivered. If this fuels some research interest in you, the topic can be studied in different cultures and languages. An existing database of speeches given by British politicians during Brexit campaigns can also be used.

Prerequisites:

• Good command of Praat or another speech processing software
• Basic knowledge of statistical inference

Literature:

Rosenberg, A., & Hirschberg, J. (2009). Charisma perception from text and speech. Speech Communication, 51(7), 640–655.

Alternative questions like Is the baby awake or asleep? are realized using a wealth of different surface cues in different languages –prosodic, morpho-syntactic, lexical– and have special semantic and pragmatic properties. This makes the mapping from surface form to utterance meaning in this construction particularly interesting from a cross-linguistic point of view.

• How are alternative questions realized in less studied languages?
• What is their distribution in embedded environments?
• What is their distribution in matrix environments, i.e., do they have special discourse restrictions?

There are many ways to ask one and the same question. Compare: (i) Is Amy at home?, (ii) Is Amy not at home?, (iii) Isn’t Amy at home?, (iv) Amy is at home?, (v) Amy is at home, isn’t she?. While some of these forms are neutral, others express some degree of bias towards or against the prejacent proposition ‘that Amy is at home’.

• How exactly do these forms differ from each other in terms of semantic/pragmatic behavior?
• What impact does prosody have in polar questions like (i)-(iii)? [Together with Prof. Bettina Braun]
• What readings do Rising Declaratives like (iv) allow in your native language?
• Besides a tag question form like (v), does your native language have other tag question forms? What are their semantic-pragmatic properties?

Crosslinguistically, the meaning of lexical quantifiers –e.g., ‘every’, ‘some’, ‘no’, ‘most’,… – is known to obey a number of mathematical properties, including extension, isomorphism and conservativity. Recent efforts in Artificial Intelligence, Philosophy and Linguistics strive to derive these universal properties from (i) learnability considerations (using neural networks), (ii) the simplicity/informativeness trade-off and (iii) properties of logical operators and/or of natural language. For each such universal property:

• What explanation is best?
• Can several competing explanations be tested experimentally?
• If a particular reading of a specific quantifier in your native language seems to violate that universal property, can this reading be derived otherwise within the grammar of that particular language?

Languages differ in how their tense and aspect paradigms are structured and what meaning distinctions they convey. For a given language, one can develop a grammar fragment on tense or aspect in LFG/XLE.

Expressions like German schon and bloss or English totally and even, etc. often live a double life in the grammar: as adverbial elements contributing to the propositional content of the sentence (e.g., in The glass is totally full) or as speech acts modifiers fine-tuning the illocutionary act performed (e.g., in You should totally click on that link).

• Does the speech act use have any syntactic/semantic/pragmatic distribution requirements? If so, what is the underlying motivation for them?
• Can we pin down the content of the speech act reading with the help of experimental methodology?
• How should the speech act reading be theoretically modelled?
• How do the two readings –the propositional meaning and the speech act meaning– relate to each other?

Different attitude verbs –e.g. think, know, wonder— select for different types of complement clauses –e.g., that-clauses vs. interrogative clauses, indicative vs. subjunctive clauses, only V-final clauses vs. also V2 clauses. What guides this selection in each particular language? And can general semantic properties be identified that guide this selection cross-linguistically?

What happens with the meaning of functional items –e.g., pronominal reference, definite and indefinite article, tense/aspect/mood morphology—in a multilingual setting? Do we observe transfer in their acquisition? Is there delay or acceleration in the acquisition process?

If you have some ideas or interest on any other topic within semantics and/or pragmatics, feel free to come to my office hours.

In this thesis, you will be conducting corpus-based research on the multilingual usage of Mary, Queen of Scots (1542-1587). Fluently multilingual from birth, Mary’s letters are written in French, Scots and English, and the aim of the thesis is to establish patterns and regularities in this usage, as well as potential cross-linguistic influences.
Required skills: Corpus linguistics, ideally experience working with historical texts, some knowledge of French and English
Language of the thesis: English or German

In this thesis, you will investigate the grammatical effects of lexical borrowings from French and Latin into English, with a specific focus on French and Latin verbs with non-finite complements. You’ll be testing the idea that these borrowings entered the system at a crucial time to trigger wider changes in complementation patterns.
Required skills: Corpus linguistics, ideally experience working with historical texts, basic syntax
Language of the thesis: English or German

In this thesis, you will investigate the Transylvanian Saxon variety of German, a language island in northern Romania. You will use the Audioatlas Siebenbürgisch-Sächsischer Dialekte to investigate some aspect of contact effects of Romanian and/or standard German on this variety and its geographical patterning.
Required skills: High level of German, ideally experience in working with recorded spoken data
Language of the thesis: English or German

You can come to me with ideas about projects with a historical or syntactic dimension to them, especially those that use corpus evidence.
Language of the thesis: English or German