Content-to-Capacity Shift

Principle

The engineering target shifts from content (what the person wants) to capacity (whether they can currently access what they want). This makes the problem tractable and avoids the authentic desire problem.

Derivation

Frankfurt (1971) distinguishes persons from wantons by the capacity for second-order volitions -- wanting certain desires to be effective as one's will. Bratman (2007) extends this with self-governing policies organized across time. "Capacity to access evaluative processes" is an engineering translation of this philosophical machinery.

The stress-induced PFC degradation described in the biological mechanisms (Arnsten 2009, 2015) specifically disrupts evaluative alignment. The wanting system (mesolimbic DA, large and robust) and the liking system (opioid/endocannabinoid hotspots, small and fragile, ~10% of NAcc volume) can diverge. The PFC evaluation that keeps them aligned is the first casualty of catecholamine disruption.

The result is the wanton state (Frankfurt 1971) stated mechanistically: under stress, the human loses not the capacity to want, but the capacity to evaluate whether what they want aligns with what they endorse.

Engineering implication

The machine monitors whether second-order governance is online -- whether the person can evaluate and endorse (or reject) their current desires. It does not need to determine which desires are authentic. This shifts the target from an impossible problem (knowing what someone truly wants) to a tractable one (detecting when their evaluative capacity is degraded).

The neurobiological mechanism for this detection is the catecholamine-PFC balance: when PFC is on the ascending limb of the inverted-U, evaluative governance is functional. When past the apex, it degrades. The measurement stack must resolve this distinction.