The building blocks of cognition: core debates in infancy research


Infant cognition is a flourishing and respected field of inquiry, but its practitioners agree that a large share of experiments are underpowered and that replicability is a serious issue (Frank et al., 2017). A consortium consisting of all four babylabs in the Dutch Baby Brain and Cognition Network will replicate two exemplary studies, each fundamental to the core debates in the field. The algebraic rule learning study reported in Science by Marcus, Vijaya, Bandi and Vishton (1999) plays a key role in the debate on (innate) learning mechanisms involved in language acquisition, and whether these are human- and language-specific, while the study on cognitive gains of bilingualism, reported in PNAS by Kovács and Mehler (2004) is of fundamental importance in the debate on whether bilingualism has a positive influence on cognitive flexibility. By replicating each study in all four labs, we will improve statistical power of the studies, and test the robustness of the original results when the experiments are performed in different labs. In our data analyses, we will depart from the traditional frequentist methods used in the original studies. Specifically, a Bayesian sequential testing approach will be used in order to determine when enough evidence is gathered to evaluate the original findings. This approach also allows us to quantify evidence for the null hypothesis in case we cannot replicate the results of the original studies. Summarizing, by increasing statistical power, the aim of the current project - led by the Dutch Baby Brain and Cognition Network - is to solidify the results of two infant studies addressing fundamental questions concerning human cognition and its ontogenesis.





Prof. dr. C.C. Levelt

Verbonden aan

Universiteit Leiden, Faculteit der Geesteswetenschappen, Leiden University Centre for Linguistics (LUCL)


01/01/2019 tot 30/06/2020