Productive Incoherence Distinction
Principle
Not all coherence reduction is harmful. Growth, learning, and developmental transitions require temporary, bounded reduction in integration -- the productive discomfort of working through genuinely hard problems. The system's target is minimizing chronic, depleting incoherence while preserving capacity for productive, bounded incoherence.
Derivation
The distinction is already present at the mechanism level. Aston-Jones & Cohen (2005) identify the need to distinguish adaptive tonic LC shifts (person appropriately exploring alternatives) from pathological tonic states (person unable to engage). The ascending limb of the Arnsten inverted-U -- where moderate catecholamine levels enhance PFC function rather than degrade it -- provides additional grounding: stress that keeps the person on the ascending limb is biologically different from stress that pushes past the apex.
A person working through a genuinely hard problem -- a career transition, a values reconfiguration, a difficult creative challenge -- may show physiological and behavioral markers consistent with the green-fragile regime (G1) while engaged in exactly the kind of growth the architecture exists to enable.
Candidate distinguishing criteria (hypothesis-level)
Three proposed operational markers, not yet grounded in specific empirical findings:
- Whether the person has awareness and agency over the tension
- Whether the trajectory moves toward resolution or is chronically stuck
- Whether the cost pattern is acute/bounded or chronic/depleting
Engineering implication
The U (uncertain) regime provides conservative protection by defaulting to null when the productive/depleting distinction cannot be made. The understanding layer may eventually resolve this distinction: an AI with access to the person's values, current challenges, and engagement history may be able to estimate whether current stress is associated with productive work or chronic depleting cycling. This is a specific instance of P4's claim that the understanding layer adds value beyond sensor-plus-rules.
The coherence-agency bootstrap (see models) provides additional motivation: a person in a productive bootstrap cycle -- where aligned action is feeding back into deeper integration -- is in exactly the state the system should protect, not interrupt.