This paper introduces selective recursion, an architectural pattern for managing self-reference and incompleteness in expressive systems. Rather than attempting to resolve Gödelian incompleteness or suppress recursive reasoning, the approach treats incompleteness as an environmental constraint and focuses on containment. The core contribution is the explicit distinction between recursive domains, which support self-referential reasoning, optimization, and critique, and non-recursive domains, which provide jurisdictionally protected axioms, constraints, or commitments that recursive processes may interpret but not revise. By enforcing asymmetric authority between these domains, systems can remain expressive without permitting recursive processes to redefine their own conditions of validity. The framework is domain-agnostic and is illustrated through schematic examples in artificial intelligence systems, legal and institutional design, and language and symbolic systems. The paper does not propose new logical theorems or normative axioms, but instead offers a general architectural strategy for preserving coherence and survivability in systems subject to self-reference.
Benjamin Ortlip (Sun,) studied this question.