We have noted the difficulties of deepening the relations between the social and functional diversity, starting when we wonder what social factors are required to describe the colloquial. We take into account, in descending order, the «associated identities» that Block (2014) analyzes: linguistic capital, cultural capital, age, gender, usual activity, «habitus / ideology / attitudes», «social class», ethnicity and religion. In this conceptual context of uncertainty, it is not surprising that the social variation dimension is inadequately treated in the Catalan prescriptive corpus: Badia (1994) confused the terms diastratic and diaphasic, and reading the Catalan Academy?s corpus (IEC) an incoherent, contradictory, inaccurate and incomplete treatment of social variation is found. The accepted conception ?prescriptive grammar?, which is based on social judgments, it is contradicted by «treatment of variation» of GIEC (2016: XXII), which excludes them. Given this contradiction, we present an analysis of how the IEC exposes social variation, functional variation and the relationships between them. Finally, we start the debate on what would be the appropriate treatment of social variation by the IEC.