09.12.2025 Olivier Rüst

Linguistic systems exhibit vast differences across the world. All children, however, acquire these with apparent ease. How can children learn such hugely diverse grammatical systems relying
on essentially the same cognitive mechanisms? I address this question using state-of-the-art computational methods.
I present three studies that tackle this question from different perspectives. I start off with small-scale speech patterns concerning the acquisition of case. Case marking is an interesting grammatical phenomenon as it marks the hierarchical relationship of constituents. I then take a step back to holistically investigate the development of child speech patterns. Here, I aim to show how a fully creative language can emerge purely based on the probabilistic combinatoriality of discrete linguistic units. Lastly, I contextualize the acquisition of speech patterns in its social environment. Specifically, this project investigates whether children are more likely to learn speech patterns during linguistic interactions with caregivers rather than purely overhearing adult speech. These projects highlight the power of distributional learning across diverse linguistic contexts, particularly emphasizing the emergence of case systems and the development of fully productive language. I also highlight the inherently social nature of language acquisition, by showing that children are more likely to learn language in interaction with their caregiver(s) rather than by passively overhearing adult language.