Tid: Torsdag 26 oktober 2017, kl. 15:00–17:00
Plats: C307, Södra huset, Frescati

It is often assumed within evolutionary linguistics that languages change in response to socioecological pressures. Language complexity is a common parameter to test for such adaptation. Strong claims have been made about the evolution of complexity. e.g. that large proportion of non-native speakers in a population facilitates morphological simplification. The evidence in support of this hypothesis comes mostly from typological data, and it is not entirely conclusive. Moreover, the typological data do not give us direct insights into potential mechanisms of simplification, and thus have to be complemented by other approaches. I report the results of several experimental and computational studies that test various aspects of this hypothesis, focusing on different facets of complexity: overspecification, irregularity and redundancy. One study uses artificial language learning to test whether lower levels of overspecification and irregularity make languages more learnable for non-native speakers. Another two use iterated artificial language learning to trace the trajectories of overspecification and irregularity under normal and imperfect learning conditions. Finally, the last study uses a novel computational method to measure redundancy and test whether redundant morphological features are more likely to be lost.

 Varmt välkomna!
Francesca Di Garbo & Richard Kowalik