Antipoietic Selection: The Missing Layer Between Systemic Self-Preservation and Civilizational Becoming

Complex systems cannot act without selection. Institutions select research proposals, risks, policies, investments, credentials, candidates, and increasingly the ranked outputs of artificial intelligence. Selection is therefore not merely an administrative operation but a threshold through which some possibilities acquire material consequence while others remain latent. This conceptual research article introduces antipoietic selection as a specific and falsifiable ontocybernetic mechanism: the recursive preservation of selection conditions that disadvantage emergent possibilities insofar as their realization would reduce dependence on the mediation, jurisdiction, symbolic labor, or gatekeeping through which the selecting system reproduces itself. The theory distinguishes antipoietic selection from structural inertia, path dependence, resource dependence, institutional isomorphism, exploration–exploitation bias, cumulative advantage, system justification, and related accounts. Its central explanatory variable is displacement potential: the degree to which a possibility, if actualized, would reduce the future necessity of incumbent mediation. The paper specifies observable mechanisms, boundary conditions, falsification criteria, operational variables, testable propositions, and empirical designs capable of determining whether displacement potential adds explanatory power beyond novelty, uncertainty, quality, feasibility, risk, and institutional fit. Applied to artificial intelligence, the framework identifies a risk deeper than automated bias: automated inheritance, through which historically stabilized criteria of relevance become invisible infrastructure for future admissibility. Sapiopoiesis is developed as a constructive counter-architecture of source integrity, corrigible selection, reversible mediation, protected exploration, and orientation before execution.