ABSTRACT Many biological and chemical systems have the capacity to change autonomously by shaping their use of environmental inputs. To better understand this capacity, researchers have developed models of goal‐directedness. This concept mainly describes a system's ability to maintain itself in the face of perturbations. Such a definition, however, fails to encompass the type of autonomy these researchers were fascinated with in the first place: that of systems which not only maintain but also transform themselves. In this paper, we start from the definition of autonomy as a system's capacity to generate the constraints which, in turn, shape this system's constraint‐generating behaviour. From there, we propose a new way to formalize autonomous change, using the framework of chemical organization theory (COT) as a modeling language. We develop our model by extending the concept of closure of constraints to systems that generate new constraints. This formalization produces a renewed, more complex understanding of autonomy and its relations with previous models of goal‐directedness. First, the definition of autonomy shifts from maintaining a singular main goal to the dynamic coordination of a multiplicity of evolving goals and processes. Second, a self‐organized, goal‐directed system can now be understood as a static version of an autonomous system. Finally, we discuss the implications of this framework for human autonomy.
Busseniers et al. (2026) studied this question.