There is little contact between theories of individual word identification and sentence processing models. In this talk, I present a study, led by Katarina Marjanovic and carried out in collaboration with Yamil Vidal, where we tried to bridge this gap by assessing cross-word priming during natural reading. By combining eye tracking and EEG, we found that cross-word semantic priming (“…the cat and the dog…” vs. “…the table and the dog…”) is solid and strong, both in eye tracking metrics and in Fixated-Related Potentials (FRPs). This is not the case for morphological priming (“…the cat and the dog…” vs. “…the cats and the dog…”), which doesn’t seem to emerge in sentence context.