Exigence and the Cognitive Kernel

Computational systems routinely model tasks, priorities, deadlines, permissions, attention, and action, but they rarely possess a unified account of why a particular state presently calls for transition. This paper introduces Exigence as a first-class computational primitive: the graph-derived condition by which an object, process, agent, institution, or environment presently calls for some admissible change. Exigence is neither priority nor urgency. Priority is declared ordering; urgency concerns time sensitivity; Exigence arises from the live relationship between state, dependency, consequence, opportunity, authority, capability, care, safety, and constraint. The paper situates Exigence within a broader Cognitive Kernel, or CogK, that separates observation, salience, epistemic interpretation, Exigence, standing, attention, custody, sufficiency, policy, authorization, execution, receipt, and outcome. These elements are treated not as interchangeable stages but as distinct projections over shared governed reality. The separation allows systems to distinguish what is observed from what is believed, what matters from what may be done, and what may be done from what actually occurred. The resulting architecture supports lawful refusal, receipt-before-effect, context compilation, agent orchestration, institutional modeling, attention governance, household coordination, autonomous systems, games, simulated societies, and the progressive crystallization of expensive semantic reasoning into deterministic machinery. It also permits the modeling of pathological Exigence topologies such as fixation, suppression, inflation, avoidance, fragmentation, and over-propagation. These structures need not be used to diagnose humans. They can instead serve as integrity and execution-gate models for detecting recurrent computational trajectories that should be refused, escalated, reinterpreted, or repaired. The central claim is that Exigence is a candidate connective layer between a passive world model and governed transition. Its value depends on keeping an Exigence claim distinct from observation, salience, standing, attention, policy, authorization, effect, and outcome. The Cognitive Kernel thereby offers a general substrate for systems that must decide what currently matters, who holds the next move, what conditions are sufficient, what transitions are lawful, and when action must stop.