From ARQUÍA to 0-1-2-3: Genealogy, Formalization, and Runtime-Assurance Implications of a Relational Preservation Framework

ARQUÍA did not begin as a universal theory of ethics. It began as an engineering and governance problem: how to allow an artificial agent to act usefully without turning optimization, obedience, self-preservation, or declared good intentions into absolute objectives. Repeated analysis of high-stakes decisions exposed a recurring structure: preserve valid continuity; preserve the parts whose continuities constitute the system; admit loss only when it is genuinely unavoidable, minimal, non-arbitrary, and proportionate; and continue searching recursively for less destructive alternatives. That structure was subsequently abstracted as the 0-1-2-3 framework and developed in ARQUÍA v7.0-NRE as the 0-1-2-3 Theory of the Relational Birth of Ethics. This companion paper reconstructs that genealogy and translates the theory back into an engineering proposal for runtime assurance. It introduces a dynamic-autonomy hypothesis: the maximum operational autonomy of an artificial agent should depend not only on its competence, but on its verified capacity to detect, contain, and reverse the adverse consequences of its actions relative to potential harm and uncertainty. The paper relates this proposal to Runtime Assurance, Simplex architectures, Control Barrier Functions, and the NIST AI Risk Management Framework, while explicitly distinguishing ARQUÍA's relational claims from established safety-engineering results. It concludes with six falsifiable hypotheses and a comparative experimental protocol. The paper is presented as a conceptual and formal research contribution open to criticism, implementation, reproduction, comparison, and falsification. It does not claim peer review or completed empirical validation.