The Reflective-Ethical Engine: A Constraint-Based Framework for Viable Cognition Under Uncertainty

This paper proposes the Reflective–Ethical Engine (REE) as a constraint-based framework demonstrating that five axiomatic constraints on viable cognition can be satisfied simultaneously within a single minimal structure. Cognition is defined as the maintenance of coherent input–output behaviour across time under permanent uncertainty. The framework rejects privileged access to ground truth, fixed optimisation targets, and rule-based ethics. Instead, it characterises viable cognition in terms of three irreducible functional constraints: fast predictive generation, deep temporal synthesis, and constrained trajectory selection. Predictive trajectories are retained or discarded based on temporal coherence and survivability under action rather than correspondence to truth. Ethical stability emerges as a necessary bias toward futures that preserve long-horizon coordination across self and others—here termed love (long-horizon coherence bias). The framework is minimal, substrate independent, and applicable across organisational scales. It is advanced not as a theory of truth or optimality, but as a constraint theory specifying what cognition cannot violate and remain viable.