Katja Stärk


I am a PhD student in the Language Development Department. For my thesis, I am using corpus, production and (potentially) EEG data to investigate statistical learning and word segmentation.



I am a psycholinguist primarily interested in first language acquisition. I am using corpus analyses and experiments to investigate word segmentation and statistical learning as well as the acquisition of morphology. I also enjoyed working on bilingualism and numeral systems.

You can find some more information about my PhD project as well as other projects below. Please, do not hesitate to Katja.Staerk [at] mpi.nl (contact me) if you want to know more about the projects.


PhD Project

Tracking statistical learning of naturalistic input

Under supervision of Rebecca Frost and Evan Kidd

One of the first puzzles children are confronted with during language acquisition is to identify the words in the speech stream. One cue they can draw on to solve this puzzle is the statistical information contained in the speech stream. In my PhD project, I am investigating the statistical learning mechanism in German. To do so, I am applying different methods: I am using a corpus analysis to look at the transitional probabilities in natural German, and I am using behavioural as well as electrophysiological techniques to study infants’ and adults’ learning. Finally, I am comparing different levels of naturalness of the input language to improve our understanding of the extent to which statistical learning contributes to language acquisition.

☞ See the OSF page about our corpus study on German CDS here.


Other Projects

Acquiring the non-existence of a morphological rule

With Jenny Tippmann, Miriam Ebersberg, Andreas Opitz and Sonja Rossi

In languages with irregular verb forms the vowel of a verb might change in the past tense form, as in English sing-sang-sung. Interestingly, not all theoretically possible rules are applied in the languages we know. For example, no English verb follows the rule sing-sang-sing, i.e. a change of the vowel for the preterite but not for the participle. Adults show different EEG responses to correctly applied rules, incorrectly applied rules and non-existent rules applied. In this project, we are looking at the developmental trajectory of children acquiring the non-existence of a rule looking at 6-7 and 8-9 year olds’ EEG responses.

☞ See a poster about the project here.

The evolution of the diverse Uto-Aztecan numeral systems

With Andrea Bender

Uto-Aztecan languages are mostly spoken on the west coast of North and Central America. Even though all languages evolved from the same (linguistic) ancestor with a single numeral system they nowadays show a vast diversity in numeral systems. In this project, we are gathering information about the individual number systems of the Uto-Aztecan languages, and modelling when and how their number systems became so diverse.



Winter 2014/2015

"Repetitorium" for Grammar Theory and Linguistic Typology
MA Linguistics, University of Leipzig, Germany


Teaching Assistant

Winter 2012/2013 –

Logic for Linguists
BA Linguistics, University of Leipzig, Germany

Summer 2013 –

Formal Semantics
BA Linguistics, University of Leipzig, Germany




Since 09/2017

PhD Student
MPI for Psycholinguistics, Nijmegen, The Netherlands

10/2013 – 08/2016

M.A. Linguistics
University of Leipzig, Germany

10/2010 – 09/2013

B.A. Linguistics
University of Leipzig, Germany


Work Experience

04/2016 – 08/2017

Research Assistant / Associate
University of Bergen, Norway


☞ See my full CV here.

Find me elsewhere

      ORCID       OSF       Github      

Share this page