Wh-word acquisition in Czech: Exploring the growing trees hypothesis

Radek Šimík, Klára Matiasovitsová, Filip Smolík

Abstract: We investigate the acquisition of different functions of wh-words in Czech (Slavic) - a language where wh-words are highly polyfunctional. Our working hypothesis is that the acquisition proceeds in accordance with the growth of the syntactic tree (aka the growing trees hypothesis; Friedman et al. 2021), which in turn is matched by the growth of the semantic representation. More specifically, children are expected to first acquire wh-interrogatives, then conditionals, correlatives, and free relatives, and finally (light-)headed relatives. The working hypothesis is set against the background of what could be considered the null hypothesis, namely that children’s production matches the adult input. We have examined the Czech CHILDES corpus (Chromá et al. 2024; 6 children 1;07-3;09), tagging all wh-occurrences (nearly 11 thousand tokens) for the construction they appear in. Using both descriptive and inferential statistical methods, we have found evidence either supporting or being compatible with our hypothesis. At the same time, our results are not consistent with a simple input-based account. We also discuss some limitations of a corpus-based approach for addressing our research question. The most significant hurdle is the overall low frequency of non-interrogative wh-words.