Exploring the Neural Representational Space of Abstract Concepts

Abstract concepts constitute a fundamental part of language, without which humans would not be able to express complex ideas, societal norms, scientific notions, their feelings and intentions. Nonetheless, they have been largely neglected in the neuroscientific literature, so that their representation in the human brain remains relatively uncharted compared to the extensive investigation that concrete concepts received, leaving many questions still open. A major debate concerns whether abstract concepts are represented only in an amodal, linguistic format, or whether they are also grounded in sensorimotor, affective and social experiential features. A second open question concerns the categorisation of abstract concepts: although abstract concepts have been treated as a unitary domain, a growing body of evidence suggests that, based on their content, they can be differentiated into distinct categories relying on separate neural bases. In this thesis, I addressed these issues using neuropsychological evidence, transcranial magnetic stimulation (TMS) and magnetoencephalography (MEG) methods. The results support a dynamic and context-dependent view of abstract concepts embodiment, whereby experiential features, either socio-affective or sensorimotor, are differentially recruited for abstract concepts representation depending on current requirements. Furthermore, while mixed evidence emerged in favour of a distinction of social and emotion concepts as two separate abstract categories, results also show that other abstract categories defined on the basis of classic taxonomy are not reflected in the brain representational space of abstract concepts. Collectively, these findings suggest that more refined theoretical and methodological approaches are needed to capture the partitioning of the abstract semantic space, and indicate that abstract concepts are flexibly grounded in experiential information.