Interests
Gender, semantics, and the mental lexicon
My work on gender examines how listeners process and interpret role nouns, pronouns, and other expressions that implicitly or explicitly encode gender. In German, this involves questions about how listeners process supposedly gender-neutral generic masculines; in English, it concerns how they is understood in singular contexts. I use computational approaches – most prominently naive and linear discriminative learning – to trace how linguistic experience shapes these patterns. This line of research brings together structural semantics, social meaning, and computational modelling to better understand how gender is resolved in real time.
Collaborators: Janina Esser, Jennifer Keller, Anita Körner, Joanna Kullik, Carolin Müller-Spitzer, Samira Ochs, Jan Oliver Rüdiger, Viktoria Schneider
Phonetics, morphology, and subphonemic cues
A recurring theme across my work is that fine phonetic detail is deeply intertwined with morphological structure. Studies on word-final /s/ in English and German show that even minimal acoustic differences can reliably distinguish morphological categories and influence both comprehension and production. Related experiments investigate whether such subphonemic variation can guide learners in artificial languages, asking whether morphological structure is not only produced but also acquired through acoustic regularities. Further work extends these questions to typing behaviour, exploring how category-specific patterns in word-final /s/ surface even in keystroke dynamics. Across these strands, I treat subtle acoustic variation as a meaningful signal that connects form and meaning.
Collaborators: Dinah Baer-Henney, Julia Muschalik, Ingo Plag
Stress, prominence, and compound structure
Together with collaborators, I explore how English compound stress arises from associations that learners form between semantic relations, morphological patterns, and prosodic trajectories. Using the framework of the discriminative lexicon model, this research shows that stress patterns do not need to be stipulated as categorical rules. Instead, they emerge gradually from mappings between meaning and experience with acoustic realisation. A complementary strand examines how listeners exploit these learned patterns during comprehension, revealing stress as a gradient, experience-driven cue rather than a purely structural property.
Collaborators: Melanie J. Bell, Ingo Plag
Sound symbolism and multimodal interactions
Another area of interest concerns how phonetic form interacts with perceptual dimensions such as shape, roundness, or spikiness. While sound symbolism has been documented extensively, interactions between auditory and visual dimensions remain comparatively underexplored. Current work investigates how such multimodal interactions influence judgements of form-meaning correspondences, contributing to a broader understanding of how humans integrate different sensory cues when interpreting linguistic material.
Collaborators: Defne Cicek, Anh Kim Nguyen, Daniel Rottleb
Computational and quantitative approaches to linguistic structure
Much of my research is united by a commitment to computational and quantitative modelling. I work with discriminative learning models, distributional semantic spaces, and context-sensitive embeddings to examine how linguistic experience shapes lexical knowledge. In parallel, I use acoustic analysis, psycholinguistic experimentation, and statistical modelling to probe the interface between linguistic structure and its realisation in speech. These tools are not treated as opaque technologies, but as methodological lenses that allow us to investigate how meaning and form become linked across the lexicon.