Author: Vladisav Jovanović
Status: Preprint
Version: Latest archived (Feb 2026)
Coherence has become cheap. Institutions, platforms, and generative AI can now produce fluent explanation and “wisdom-shaped” discourse without paying the costs that historically tethered coherence to reality. This paper argues that a central failure mode of the AI era is answerability loss: systems optimize for coherence while filtering constraint, delaying consequence, and suppressing revision. I introduce Structural Intelligence (SI)—at the intersection of Philosophy of Mind, Epistemology, and Systems Theory—defining intelligence not as coherence-production but as the organization of form under pressure: the capacity of an agent (human, institutional, or artificial) to remain coherent and corrigible under activation, incentives, and time. SI distinguishes coherence (internal intelligibility) from contact (constraint-coupled truth that binds through cost, revision, and consequence) and proposes a five-pillar Structural Audit (Honesty, Action, Resilience, Connection, Self-Awareness) modeled as a tensegrity structure governed by three feedback loops (Sovereignty, Metabolic, Answerability). Two mechanisms sharpen the criterion: maintenance inflation (the rising cost of defended coherence) and the invariance constraint (a hard floor separating sovereign agency from signal-optimization). The paper closes with a minimal audit protocol for AI governance, institutional integrity, and personal agency.
structural intelligence; agency; sovereignty; cheap coherence; reality-contact; presence; answerability; constraint; revision; falsifiers; binding