Abstract

This paper applies the Coherence Engine (CE) framework, a general architectural theory of consciousness developed through phenomenological observation and described in full elsewhere (Sowell, 2026a), to the domain of mental health dysfunction and recovery. The CE framework proposes that consciousness is produced by five functional modules: the Prediction Engine, the Orchestrator, the Gating System, the Error Integration Loop, and the Core Self Modulator (CSM), operating together to maintain a unified, prediction-generating, self-modeling subject. This paper argues that mental health dysfunction is not a collection of discrete biological disorders but a set of identifiable architectural failures, each occurring at a specific depth and through a specific mechanism. The framework is engaged with the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) throughout: where the DSM’s symptom-cluster descriptions correspond to architectural predictions, that correspondence is noted; where the architectural account reveals mechanisms the DSM does not explain or reframes conditions the DSM pathologizes, those divergences are examined directly. The paper develops architectural accounts of mood disorders, anxiety, trauma and PTSD, obsessive-compulsive disorder, personality disorders, dissociative identity disorder, reactive attachment disorder, autism spectrum condition, addiction, schizophrenia spectrum conditions, attention deficit hyperactivity disorder, eating disorders, grief, sleep disorders, substance-induced states and withdrawal, and somatic symptom disorders. Autism and ADHD are reframed as Gating System calibration variants rather than deficits. A depth-of-compilation framework is developed to explain why different interventions reach different conditions, and why insight frequently fails to produce durable change. The paper concludes with a unified account of what recovery consists of at the architectural level, and what conditions make it possible.