Wright–Wrong: A Treatise on Decision Guidance Under Constraint

This book examines how decisions labelled “right” or “wrong” emerge in complex human and artificial systems, without relying on moral absolutes or ethical doctrine. Drawing on familiar ideas from statistics, optimisation, game theory, systems dynamics, and topology, it shows how different ways of modelling the same situation naturally produce different, yet defensible, conclusions. The central claim is that “rightness” is not a fixed property of the world, nor a prospective truth that can be reliably located in advance. It is better understood as a retrospective stabilisation signal — a local, conditional indication that a system has not (yet) destabilised under constraint. Moral language, governance processes, and even formal models are analysed as coping mechanisms near unobservable boundaries (event horizons of observability), where prospective correctness fails and parameter-freezing becomes necessary for action. The text refuses normative rescue operations or new ethical frameworks. Appendices provide spreadsheet-reproducible models that demonstrate why lenses (distributional, feasible-region, equilibrium, dynamical, topological, fuzzy, Bayesian) disagree structurally — not from error, but from geometry. This matters increasingly in an age of algorithmic decision-making and AI-assisted governance, where explicitness collapses classical moral opacity. This deposit is version 1.0 (pre-final / archival preprint). Future versions may include minor clarifications, expanded appendices, or formatting adjustments.