Conditions for Coexistence - Structural Notes on Governing Human–Artificial Systems

Abstract Most contemporary discussion of artificial intelligence governance relies on at least one stabilising fiction: moral alignment, benevolent intent, human-like understanding, or the eventual convergence of values. This work rejects those premises. Instead, it treats artificial systems as participants in hybrid human–machine environments characterised by asymmetries of speed, consequence, persistence, and interpretability. Within such environments, stability does not arise from intention or ethics, but from constraint, corrigibility, and the management of surplus capability under bounded authority. This paper develops a structural vocabulary for reasoning about coexistence between humans and artificial systems without anthropomorphism. It distinguishes behavioural success from participation in consequence, explanation from control, and recognition from restraint. Governance is analysed not as moral arbitration but as the allocation and containment of power within systems subject to drift, scale, and incomplete formalisation. The focus is deliberately narrow. Questions of consciousness, alignment, and ultimate moral standing are bracketed as structurally irrelevant to the near-term stability of hybrid systems. What matters instead are scope, sanctionability, predictability, and the preservation of reversibility under pressure. This preprint presents the conceptual foundations and constraints that underpin the larger work Conditions for Coexistence. Later sections, not included here, examine failure modes that emerge when behavioural fluency outpaces institutional capacity for control.This work applies frameworks developed in RightWrong and Models That Break When Taken Seriously, and structural analysis developed in On Stability, Structure and Consequence.2026-03-06:production version