Many complex systems generate candidate states, evaluate those states against constraints, and continue operation only while those constraints remain satisfied. This paper formalises this pattern as a recursive admissibility architecture within the Paton System. The framework demonstrates how internal structural relations (Tier-2), admissibility constraints (Tier-3), observable outputs (Tier-4), and recursive continuation (Tier-5) together produce the operational behaviour observed in real-world systems. Neural networks and other learning systems provide a clear example of this architecture in practice. The result establishes a structural bridge between the internal architecture of systems and their observable behaviour across domains.
Andrew John Paton (Sun,) studied this question.