Coherence Before Optimization

Modern systems across domains are increasingly designed around optimization—performance, reward, efficiency, or growth. While optimization is powerful, systems that prioritize it without an explicit account of coherence frequently exhibit drift, fragility, or sudden collapse. This essay advances a structural reframing: coherence must precede optimization. Coherence is defined as the preservation of identity-defining invariants under change. Optimization is not rejected, but repositioned as a secondary activity that must remain bounded by coherence. The contribution is conceptual and architectural. No algorithms, implementations, or domain-specific prescriptions are proposed.