Our team’s long-term project investigating social cognition in a range of languages is based on a corpus typology approach. We use a set of annotation schemata to code particular instances of language use that we see as indicating some aspect of social cognition. We then compare the amounts and types of instances by language or task participant. Much careful consideration went into designing annotation schema to look at various domains of social cognition. This set of guidelines describes our eight coding schemata so that our results are interpretable and to make our scientific process open. The guidelines should also provide enough information for people who are not part of the project to code their own language data.