The social sciences can take better account of dyadic relations than they have without the need for a complete overhaul and the creation of a new paradigm. If we conceive of relations as entities that emerge from interactions, it is possible to integrate them into a robust general ontology. In the ontology proposed in this text, I distinguish relations between human entities (persons, collectives) and other entities (animals, plants, minerals, artefacts, discursive or cognitive elements) from relations between human entities. I then define the notion of collective as a set of persons sharing relations with entities. A collective can be analytic (defined from the outside) or explicit (claimed by its members). Interpersonal relations can then be defined as explicit dyadic collectives. This implies specifying how the dyadic level is specific and deserves special attention from the social sciences, which justifies the development of social network analyses. However, it also means an interest in the relations between collectives or individuals and collectives and the dynamics created between collectives and networks. Two other categories are necessary to complete a social ontology attentive to relations: the institution and the sphere of activity.