My research develops empirically grounded models of complex meaning in discourse and interaction. I investigate pragmatic and semantic phenomena that are highly context-sensitive and rely on extralinguistic knowledge, including stance, fear, (mis)trust, and common-ground negotiation. A particular focus of my work lies on the linguistic patterns through which such meanings are constructed in texts and conversations, including argumentation patterns, name-based lexical formations, and grounding acts. Methodologically, I combine human annotation of complex interpretive categories with corpus linguistics, NLP, and experimental approaches. By integrating hermeneutic and computational methods, my work connects close reading with scalable digital analysis at the intersection of linguistics and the Digital Humanities.

Discourse, media, and evaluative meaning

A central strand of my research examines how public discourse constructs complex evaluative meanings such as trust, mistrust, fear, and stance. My doctoral research analyzed the linguistic construction of trust, mistrust, and fear in German media coverage of the refugee debate and developed corpus-based approaches to the study of argumentation patterns (topoi). I have since extended this line of work to other domains of public discourse, including Euroscepticism, COVID-19, and current debates on language models and AI.

Context-sensitive meaning and extralinguistic knowledge

Another focus of my research lies on meanings whose interpretation depends heavily on individual and collective knowledge beyond the linguistic signal itself. This includes my habilitation project on name-based word formation patterns in German, such as personal-name compounds and name blends, which examines how context, discourse domain, and media environment shape the production and interpretation of these formations. More broadly, this work is concerned with how linguistic patterns activate socially shared knowledge and support the interpretation of context-sensitive meaning.

Grounding and human–machine interaction

A further strand of my research investigates grounding and common-ground processes in both human–human and human–machine interaction. Here, I examine how language models interpret and generate context-sensitive meanings and how their performance relates to users’ trust and calibration. This line of work extends my broader interest in common ground, interpretation, and the linguistic conditions under which humans and machines can successfully coordinate meaning across texts and dialogues.