Bourdieu has left his field theory unfinished. Recent publications - the lessons at the Collège de France (Sociologie générale, vol. 1 and 2) and the collection Microcosmes – offer new insights to how he was trying to develop this theory. Furthermore, empirical research done about various fields enable us to propose a new theoretical framework for rethinking the relations across fields. These relations are of three sorts: homology, dependency and exchange. The paper will include a reflection on the concept of subfield, which is undertheorized in Bourdieu’s work, but appears to be highly heuristic for constructing certain objects such as the subfield of critics, of literary agents, or of academic publishing.