F4: Controllability Inference Is Itself State-Dependent
Integration
This premise synthesizes two findings into an engineering-critical feedback loop:
- Stress biases controllability inference toward perceiving uncontrollability (Ligneul et al. 2020; fMRI evidence for mPFC encoding of controllability prediction errors).
- Computational modeling (Karvelis & Diaconescu 2024) predicts a positive feedback loop: stress biases controllability inference toward uncontrollability -> this prevents activation of the vmPFC protective pathway (F3) -> this permits further stress escalation -> which further biases controllability inference.
Why this matters for the architecture
The feedback loop is the mechanism that makes degraded states self-reinforcing. A person in a red state perceives less controllability than actually exists, which prevents the very mechanism (vmPFC-DRN activation) that would restore their state. Breaking this loop is the stabilizer's primary function.
The architecture breaks the loop by providing external controllability signals -- structured micro-interactions where the person detects action-consequence contingency -- rather than waiting for the person to find controllability on their own. This is why the red-regime actuator table allows "micro-controllability" but prohibits "high-load planning": the former provides the specific input the vmPFC-DRN pathway needs; the latter adds demands the person cannot meet.
Evidence status
SUPPORTED for the stress-induced bias. DERIVED for the escalating feedback loop (computationally modeled by Karvelis & Diaconescu 2024, not empirically demonstrated in humans in real time).