Abstract : The knowledge communities as spaces hybrid involve, first, individuals who interact within a social network that produce content locally, and on the other hand, semantic entities, related within a semantic network with an autonomous structure. This thesis aims to explore the dynamics that drive social and semantic knowledge of these communities and their structural coupling. It relies primarily on two types of empirical objects: the scientific and the political blogosphere, which we follow the activity through in vivo observation of their finger marks. Our ambition is to achieve a realistic phenomenological reconstruction sociosémantiques dynamics of these communities of knowledge. To do this, we distinguish between the dynamics associated with morphogenesis and phylogeny of the networks that make up these systems and dynamic processes of diffusion they support. The methods for characterizing the morphogenesis of knowledge communities can, first, identify and represent the structures and emergent properties of these knowledge communities, and secondly, to highlight the individual behaviors that animate regular . The phylogeny of knowledge communities involves the reconstruction of the dynamics of these high-level structures emerging during morphogenesis. The analysis of the diffusion process leading to questions about the structural determinants that affect micro and macro information flow in these communities. All these approaches allow us to describe the knowledge communities as social systems whose dimensions and semantics are, in essence, co-evolution.