Answers are grounded in this paper and cite their sources.
Investigates protocol influence on outcomes in institutional settings, suggesting implications for behavioral economics.
A programme that models institutions running on conscience has priced the levers an institution has over blame — duty, a legal safe harbour, image concern, and the administered harbour that publicises attribution in the act of capping it. A companion paper showed that on the boundary where inattention first becomes self-sustaining those prices are one linear functional on the plane of protocols, and that the structure exists exactly where the monitor’s attention cannot respond to the protocol. Where attention is destroyed it does respond; that paper found no single functional and named the gap. This paper fills it, and the answer required a more elementary fact than the question. A protocol acts through two coordinates: a floor the monitor bears whatever the audience concludes, and a coefficient on what the audience concludes. The floor and the monitor’s private production value enter his payoff through functions whose difference, in the combination given by the stake, contains no attention at all. Swapping the two at that rate therefore moves his payoff by a constant, which is invisible to every derivative, every difference, every critical point and every comparison between actions: the two are perfect substitutes at the stake, identically in the action. The coefficient has no such partner. The duty case of this is a companion note’s published decomposition, credited here; what is new is that the rate attaches to the floor alone, independently of what the coefficient does — which the duty ray cannot show, because duty moves both coordinates together. Everything follows from that asymmetry. On every boundary cut out by a derivative or a difference of the payoff, the price of a protocol change is a one-form whose first coefficient is the stake — the same number on every boundary, at every protocol and at every attention, because a substitution rate does not know which boundary it stands on — and whose second is the boundary’s own condition read at the boundary’s own attention. The boundaries at frozen attention are therefore straight lines in the substituted coordinates, and the boundary where attention is destroyed is their envelope, with the destroyed attention as the family’s parameter. The companion’s two-part result becomes one fact: the members are lines, and an envelope of non-parallel lines is not one of them. The consequences are an answer, a contrast and a threshold. The rate on the upper boundary is not the published one — the difference is exactly the second coefficient, and the two agree only where the envelope touches the member through the origin. The family of protocol changes a boundary cannot see is a straight line on the trap boundary and a curve on the upper one, so a neutral rule can be written down once there and must be read again at every step here, the gap between them being the price of not re-reading. And an administered harbour has a critical loudness for the upper boundary that is not the published one for the floor, with a band between in which its two diagnostics disagree — one seeing pure backfire and the other pure gain, and both right — while a published incidence identity for the floor does not transfer at all: with no private value the floor is inert in loudness, and the upper boundary does not exist until loudness creates it. The plane also carries two rays through the origin, one extinguishing the trap and one making engagement indestructible, and the second reproduces a threshold published elsewhere while being exactly where the price of this paper’s own stake convention falls to zero. Both rays are statements about the recovery manifold, not about any one branch of it, and the paper keeps that distinction because a companion’s threshold has already been read the other way. Magnitudes are calibration values, quoted at the house cell except where a second cell with slack is named, and each carries its stake, its protocol and its direction convention. The harbour results are stated on the algebraic domain, with the exclusion demonstrated and the loudness band inside which they also lie within the share interpretation measured rather than assumed — its lower end a threshold on the administration’s initial slope, which every technology in the class inherits, and its upper end a property of one technology, which the paper says rather than lets one symbol carry both. No welfare ranking is offered: the substitution is exactly neutral for behaviour and exactly not neutral for the monitor’s payoff, and the substrate prices neither the imposition of a floor nor the provision of private value.
No takes yet. Share an insight, caveat, or question.
Mikio Hanaeda (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: