Content and Language Integrated Learning (CLIL) has gained widespread acknowledgement in numerous education systems not only in Europe but also in other continents since its onset in the 1990s in Europe. However, despite its extensive reputation, CLIL has not yet attained a fully developed educational model. This study highlights the need for a coherent theoretical framework compatible with competency-based CLIL practices. The first part of this paper reviews the literature on CLIL, and relevant literature on foreign language teaching and sociocultural theories of learning. In the second part of the paper, the authors, drawing on the outcomes obtained from the CLIL research studies they have carried out, propose a socially mediated activity (SMA) framework for CLIL practices designed especially for young learners. In the conclusion part the authors refer to their future research projects, discuss some potential CLIL research perspectives, and suggest some research tasks that could enable researchers to develop well-informed theories and models for successful CLIL practices.