SADA Subtypes and the AAI Validity Problem: Why Earned Secure Is Not a Unitary Category — and the Default Setting Model

The Adult Attachment Interview (AAI) classifies individuals as "Earned Secure" based on narrative coherence, irrespective of developmental attachment history. This paper proposes that Earned Secure is not a unitary category at the level of autonomic neurobiology. Individuals who achieve narrative coherence from an anxious (preoccupied) base versus an avoidant (dismissing) base carry fundamentally different Periaqueductal Gray (PAG) column biases — latent under normal conditions but emergent as distinct SADA subtypes under chronic stress. Three clinically distinct subtypes are proposed: SADA-C (Caregiving), originating from a preoccupied base with vlPAG appeasement as default; SADA-P (Punitive), originating from a dismissing base with dlPAG withdrawal/punishment as default; and SADA-O (Oscillating), originating from a disorganized base with competing dlPAG/vlPAG activation producing rapid cycling — the neurobiological foundation of trauma bonding. The paper further introduces the Default Setting Model: a threshold-based framework distinguishing Low Threshold (chronic presentation) from High Threshold (High Performer collapse) profiles. The 2×3 Default Setting Matrix produces six clinically distinct presentations from a single neurobiological architecture and explains why high-functioning individuals who maintain earned-secure functioning under extreme load often exhibit the most abrupt and primitive hardware fallback when threshold is finally crossed. An Accountability Structure is specified: the Default Setting itself is hardware and outside conscious control; the threshold at which it activates is trainable through Hardware-First interventions (ECS restoration, HRV biofeedback, sleep architecture normalization); the earliest warning signals are learnable; and post-activation repair is fully within individual responsibility. Addendum A specifies implementation of the Default Setting Model as a Stress-Fallback-Profile (SFP) output layer in the ABM Neuro-Audit, including subtype-specific Earliest Warning Signal sentences for all six matrix combinations. Part of the ABM Blueprint Independent Research Series. For clinical tools and implementation: https://abm-blueprint.org