People are per definition social animals. Some even argue that the main reason the human brain is so developed is precisely to support the intricate skills required for successful interactions with others in an increasingly complex social environment. Today, our globalized society indeed requires us to adopt different modes of social interaction in different social contexts, whether this is during a meeting, down at the pub, on Facebook, or in your own household. Whereas some of the required skills are presumably innate (e.g., neonate imitation), the vast majority of such skills are learned through culture and education. In this light, our central hypothesis is that people learn to interact socially with their peers. Congruently, the project aims at identifying the mechanisms through which this learning occurs, and propose that these mechanisms are largely similar to those found to be at play in cognitive paradigms such as implicit learning. Relevant questions that will be addressed include: How do we learn to monitor and to some degree control others’ emotional state? How do we build up our “social grammars” — the informal set of rules we follow when interacting with others? Is the knowledge that subtends such “social grammars” conscious or unconscious? Specifically, we will investigate (1) to what degree people learn such social grammars unconsciously, and what happens when they are made conscious of the rules of the grammar; (2) what makes this social learning special – to what degree does it differ from other forms of learning; (3) which brain regions subtend unconscious learning of social interactions (as compared to conscious or non-social learning); (4) whether there is a difference in the degree to which unconscious or conscious social learning breaks down in pathologies such as autism and schizophrenia, in which patients are specifically impaired when it comes to social interaction.