D2: Capacity as Terminal Objective (Meta-Capability Defense)
design-argumentderivedintegrated
ClaimFor an agency-supporting system, capacity for self-directed action is the only non-contradictory terminal objective; any other objective optimizes for something the person did not choose.
This claim fails if
If autonomy-capacity is shown not to function as a meta-capability, or if a compliance-optimized system achieves equal long-term capacity outcomes.
The objective function instantiates the capability approach (Sen 1992, Nussbaum 2000) as a real-time control target, operationalizing procedural autonomy (Frankfurt 1971, Bratman 2007) as a measurable state variable.
The specific argument: autonomy-capacity functions as a meta-capability -- the precondition for any other capability mattering to the person. A person who cannot access their own evaluative processes cannot meaningfully exercise any other capability, endorse any intervention, or give informed consent.
Must be defended against three objections
- Paternalist objection: The architecture does not prohibit paternalistic intervention in extremis (health as boundary constraint). It prohibits it as the default operating mode.
- Pluralist objection: Autonomy is not the only thing that matters. It is the precondition for anything else mattering to the person. This argument is available within capability theory but is not Nussbaum's own position.
- Consent/capacity circularity: The exit-capability invariant and the removal test (E4) are specifically designed to address this.