Abstract: Speakers refer to abstract concepts such as faith, merit or norm, to convey cultural meaning and share ideas about religion, values, politics and so on. Concreteness effects i.e. processing differences between concrete and abstract words, are foundational for studies on language comprehension and memory, but they are overlooked in production studies and the associated psycholinguistic models. This is likely due to the fact that the traditional task in production studies, picture naming, is by essence restricted to imageable concrete objects. Looking at deeper semantic effects in word production thus requires a paradigmatic shift. In this talk, I will present experiments using naming from definition tasks to address these issues, as well as preliminary work on segmentation/clustering of EEG data to highlight conceptual and methodological challenges in establishing the (neuro)cognitive correlates of word retrieval.