The Boundary of the Individual as an Authoritative Causal Present — An Architectural Hypothesis

Abstract

This study analyzes the publicly disclosable requirements, prior work and evaluation plan for the functional composition B-Core = A-Core + Purple Box / Present. A-Core is a general cognitive and execution system that may also use memory, recurrence and parallel processing. Purple Box / Present is the undisclosed, history-dependent addition; its added role is a research hypothesis. The composition denotes a single B-Core instance. In the deployment examined, the private component remains in a habitat, and the system may use zero or more physical avatars. If connectivity is lost, only the A-Core functions actually deployed locally remain available; the local branch receives no permission to write directly to the central Present. The targeted narrative prior-work map connects conceptual questions of changing embodiment, distributed state authority, offline reconciliation, persistent agent memory and personal identity. A substantial proportion of the public requirements have known antecedents. A new name, a combination of partial mechanisms or an exact expression not established in this targeted review does not establish novelty. Whether the private mechanism is internally identical to or different from another mechanism cannot be determined from the disclosed functional role. The paper distinguishes identification, authority over state changes and continuity of cognitive operation. The uniqueness of committed states is a requirement, not an algorithmic property proved in the manuscript. A finite log may supply a counterexample, but by itself does not prove a general invariant. Physical event history, the accepted record and the permitted effect on the Present are separate objects of inquiry. The main empirical question is G5: what changes when the Purple Box is added to the same A-Core, and does it offer a practical advantage over the strongest technically sound alternative within the same total resource budget? The two comparisons address different estimands. During long-term closed-loop operation, subsequent memory and observations may be consequences of the intervention, so their forced identity cannot be required. The plan requires valid interventions, strong controls and prespecified endpoints. The contribution consists in clarifying compositional and permission boundaries, and organizing falsifiable subhypotheses and comparisons. The study does not present a working Purple Box implementation, a formal system proof or new experimental data. Novelty and added benefit: NOT_ESTABLISHED. The hypothesis is partially testable; full evaluation requires a concrete artifact, an external contract and a preregistered study. © 2026 Omri Bankuti. The public research document and its original figures: Creative Commons Attribution 4.0 International (CC BY 4.0). The undisclosed Purple Box / Present master: PRIVATE / All Rights Reserved. Any separately released public software or test harness: MPL-2.0. The precise scope of the three layers is specified in Sections 21–22.