Work

CONSTRAINT: the feasible-set axis of control

Year: 2026
Type: article
Abstract: The carrier-layer model of control in this series tracks instruments that move an actor's objective ($theta$) across its declared, dispositional, and operative layers. A second, equally large family of control systems does something different: it never touches the objective at all. Prisons, borders, sanctions, property regimes, antitrust decrees, agent confinement, tool-permission surfaces, and veto points do not persuade; they deform the {feasible set} $F$ --- what the actor can do, whatever it wants. This paper adds $F$ as a second axis of the control model, shows that the canonical instrument taxonomies (carrots/sticks/sermons; NATO) are structurally blind to the axis they conflate, and unifies three literatures that each hold a fragment of the idea: Sen's capability approach ($F$ at the person), constrained-MDP safety ($F$ at the policy), and access control ($F$ at the surface). The unification matters because it exposes the question the other papers in the series left open: {which} feasible-set deformations are legitimate. The answer the framework supplies --- the deployer's own $theta$, checked against whether the removed actions would have harmed the cared set --- distinguishes floors that protect life capacity from walls that protect incumbents, using the same instrument grammar on both sides.
Language: English
Open Access status: green
OpenAlex ID: W7235873181
FWCI:
Cites: 3
Cited by:
Related to: 10