External web page

Children around the globe acquire language and with it the human ability to communicate complex thoughts. This project develops a new linguistic theory to explain language and its acquisition. Our central hypothesis is that language radically compresses thought structures to sound or sign. While current theories assume a parallel between thought and language or meaning-preserving transformations, we assume that thought is mapped to language by only realizing some pieces of conceptual representations. Adult language is hyper-efficient at compressing information. For this reason, Leibniz and many others over the last 300 years have been unable to agree on the primitives of human thought.

We predict that child languages are a better mirror of the human mind. Our initial evidence suggests that children are not able to compress conceptual representations as efficiently as adults. Sometimes children produce more material than adults, leading to so-called commission errors, which have never been systematically investigated. Furthermore, comprehension is easier for children when there is a one-to-one match between language and thought. To test our central hypothesis and specify how conceptual structure is compressed into language, we carry out a series of at least twelve targeted language acquisition studies on a global scale. We have recruited collaborators for more than 50 languages from 21 different language families, two sign languages and two creoles to carry out our studies. With this data, we can formulate a complete formal model of the semantic primitives, their combination into conceptual structures, the morphological compression mechanism, and the acquisition process within our model. To accomplish these goals, we rely on insights from formal semantics, generative syntax, distributed morphology, and several other linguistic frameworks. As part of our work, we also create the first open, global research collaboration to conduct language acquisition studies.

The project involves three research groups. In addition to the Generator group at ZAS, the Compressor group working at the Humboldt-Universität zu Berlin (PI Artemis Alexiadou) and the Acquirer group working at the University of Milan Bicocca (PI Teresa Guasti).