A Matter of Time: Towards a General Theory of Agency
2026-06-22 • Artificial Intelligence
Artificial Intelligence
AI summaryⓘ
The authors explore how agency—the ability to act with purpose—emerges from living systems by focusing on how these systems are organized in time. They use a new approach that sees biological organization as a network of time-dependent processes influencing each other in an asynchronous way. This lets them clearly separate concepts like autonomy, goal-directed behavior, and agency based on how a system anticipates and interacts with its environment. Their framework connects ideas from biology and philosophy to explain a range of systems from simple chemical reactions to complex living beings. It also rethinks some existing theories about how organisms predict and respond to their surroundings.
agencyorganizational closureasynchronous dynamic Bayesian networkautonomygoal-directednessanticipationbiological organizationrelational biologyprocess ontologyphysical biosemiotics
Authors
Amahury J. López-Díaz, Carlos Gershenson
Abstract
Agency is often invoked in research on philosophy, biology, and cognitive science without a clear account of how it originates from material organization. Building on temporally parametrized (F, A)-systems, this paper develops a graded organizational theory of agency grounded in relational biology, physical biosemiotics, and process ontology. We argue that self-referential closure cannot be adequately conceived outside time: once the constitutive processes of a semantically closed organization are associated with distinct characteristic timescales, the organization unfolds into an out-of-sync dependency structure that can be formally redescribed as a history-dependent, revisable Asynchronous Dynamic Bayesian Network. This move allows for a principled distinction between autonomy, goal-directedness, agency, and open-endedness. Autonomy arises from precarious closure to efficient causation under material openness; goal-directedness from the maintenance of viability-supporting organization; agency appears when such organization acquires an endogenous anticipatory structure that selectively modulates organism-environment coupling in light of possible futures; open-endedness begins when this anticipatory organization can reconstruct its own future space of possibilities. Our framework reconciles Rosennean anticipation with organizational closure, restricts Markov blankets and active inference to derived formal redescriptions rather than first principles, and reinterprets computational enactivism in non-Fristonian terms. By deriving weaker temporalized organizations, our contribution outlines a hierarchy from proto-agential chemical systems to fully semantically closed agents, with implications for multicellular organisms, synthetic lifeforms, and neuroscience.