08.09.2023New paper on the Agent Preference and Surprisal published by IVS members Huber, Sauppe, Isasi-Isasmendi and Bickel

Eva Huber, Sebastian Sauppe,  Arrate Isasi-Isasmendi,  Ina Bornkessel-Schlesewsky, Paola Merlo, Balthasar Bickel have released a paper with the title "Surprisal from language models can predict ERPs in processing predicate-argument structures only if enriched by an Agent"

This paper revisits EEG studies on language processing with Basque, Hindi and German speakers and analyses the extent to which the agent preference and/or surprisal can predict the EEG results. It finds that both the agent preference and surprisal are needed to accurately predict the EEG results, showing that the agent preference in language processing is partially independent of linguistic experience.

Neurobiology of Language: Surprisal from language models can predict ERPs in processing predicate-argument structures only if enriched by an Agent