Abstract
Behavioral medicine has long treated medication-taking and medication adherence as the same phenomenon. This theory proposes they are not. Medication-taking is a behavior. Persistence is the capacity to adapt that behavior over time. Medication adherence is the space that emerges when that behavior is sustained. The claim is structural: it concerns what kind of phenomenon adherence is, and what kind of framework is required to understand it and act on it.
The previous version of this theory modeled behavioral systems as positions in three-dimensional state space defined by Architecture, Execution, and Context, combined multiplicatively as V = (A × E × C)1/3. Systems below V ≈ 0.70 were proposed to face elevated risk of behavioral collapse. The framework was acknowledged as entirely untested, which it remains.
This version extended that foundation by formalizing the emergent-space definition of adherence, introducing a three-state model (Resilient Stability, the Edge of Chaos, and Catastrophic Fragility), and defining predictive emergence precisely: the detection of trajectory from Resilient Stability toward the Edge of Chaos before a contextual shock triggers catastrophic transition.
Two further contributions extend the geometric framework. First, it situates TPE explicitly within the broader behavioral science landscape by identifying four structural gaps in the dominant categorical diagnostic framework that the geometric model is positioned to address. Second, it introduces a Context locus subclassification distinguishing individual, organizational, and structural origins of Context constraints. This subclassification does not alter the stability calculation but determines which category of Context intervention is actionable, and deploying the right intervention at the wrong level produces effects equivalent to targeting the wrong dimension entirely.
Seven testable predictions from the prior version are retained. An eighth prediction is added: that within Context-limited patients, interventions targeting the correct locus of the constraint (individual, organizational, or structural) produce meaningfully larger effects than interventions targeting the correct dimension at the wrong level. This framework remains entirely untested. All claims are theoretical proposals awaiting prospective empirical evaluation.
Keywords: predictive emergence, behavioral stability, medication adherence, emergent space, state space dynamics, phase transitions, critical transitions, trajectory detection, intervention timing, three-state model, geometric mean, dimensional necessity, context locus, level of analysis, intervention targeting
