This paper introduces Coherence Density as an operational framework for detecting early functional degradation in complex adaptive systems before visible collapse occurs. The central diagnostic shift is simple but significant: rather than asking whether a system recovered, it asks what recovery cost. A system may appear stable while internally paying a rising price to maintain that stability. Coherence Density is designed to make that hidden cost visible early. The framework integrates three dimensions; Informational Coupling, Phase Stability, and Energetic Efficiency, with Cost of Recovery as an independent diagnostic indicator. The hypothesis driving the framework is that Informational Coupling degrades first, before the other dimensions show visible decline, creating an early warning window that single-metric approaches miss. Computational validation using literature-grounded EEG degradation trajectories demonstrates a 10-15 unit early warning advantage over simpler alternatives under asynchronous degradation conditions. Empirical validation using the Temple University Hospital EEG Seizure Corpus is the identified next step. The framework applies across domains in neuroscience, ecology, psychology, artificial intelligence, and relational systems, because the diagnostic question it asks is domain-independent: how much internal correction is required to keep this system functioning? Presented with full acknowledgment of limitations and without overclaiming. The formula is provisional. The hypothesis is testable. The pilot is proposed.