The Generative Attributional Bond: Conceptual Foundations for a Generative Sociology

Millions of people now sustain prolonged, emotionally significant, and cognitively formative exchanges with artificial systems whose ontological status remains undetermined. The available categories (use, interaction, relationship, parasocial relationship, prosthesis, assemblage) partially capture this phenomenon but fail to define its specific structure. This paper proposes the concept of the generative attributional bond to define this specific form of sustained coupling between a human and a responsive artificial entity. Its persistence depends on asymmetric reciprocal attributions, and it produces distributed cognitive, affective, linguistic, and sociotechnical transformations. The ontological question concerning the nature of the entity is methodologically suspended; what matters are the conditions under which attributions are sustained. In this paper, “generative” is used in a sociological rather than a computational sense: it refers to the transformative effects of the bond, not to the technical capacity of generative AI systems to produce outputs. On this basis, the paper proposes generative sociology as an emerging disciplinary field, organized through three analytical planes (epistemological, technological, and innovation), and three axes of development: research, training, and public engagement.