The project aims to design and test a new, feasible strategy to combat misinformation based on naturally occurring cognitive biases, namely the similarity between misinformation and its contradiction.

This project builds on the embodied cognition view that language comprehension involves partial simulations of bodily and perceptual experiences, while addressing gaps in how visual processes integrate with sentence-level meaning. It investigates whether low-level visual processes merely interact with, or actively contribute to, semantic processing, a distinction that is crucial for theories of semantic cognition and for advancing human-like artificial language models. Using methods from experimental psychology, psycholinguistics, and corpus linguistics, the project tests how linguistic and perceptual processes mutually constrain each other through interference paradigms and fine-grained time-course analyses. It also examines individual differences in embodied comprehension and evaluates whether multimodal artificial language models outperform purely linguistic ones.