This paper introduces Transmutarianism, a novel ethical framework that evaluates moral work from transmutation: the relationship between what an agent absorbs (receives) and what it emits (provides to others) across hierarchically-weighted human need dimensions. Unlike consequentialist frameworks that treat agents symmetrically, Transmutarianism accounts for differential starting conditions. The framework's foundation is a declared taxonomy of human need dimensions; Maslow's hierarchy is its default implementation, and comparative human-needs frameworks (Max-Neef, Doyal and Gough) substitute without changing the functional. This provides both theoretical structure and empirical measurability. Transmutarianism operates as moral accounting with declared normative parameters: it measures flows of relational value using economic mechanisms, declares the weights and standards that make each placement normative, and leaves questions of moral worth to cultures, religions, and individual communities. The framework is designed to replace the practice of moral exhortation of individual agents with distributed measurement of network effects. The Conduit, the agent at the origin who passes through exactly what they receive, establishes the morally neutral baseline from which all transmutation is measured. Because deterministic passthrough is the default, the framework emphasizes systemic intervention over individual moral exhortation: designing environments where neutral processing produces net positive flows. The asymmetry coefficient serves as a cultural control mechanism, allowing different communities to calibrate the relative weight of filtering versus amplification according to their values, and the power-asymmetry coefficient holds more powerful agents to a higher standard by shifting their neutral origin up the moral axis, so that pure passthrough scores negative from power. Changes since v16, via the unpublished v17 revision (2026-06-10):- Claim 13a stated: pure passthrough by a powerful agent classifies into the Extractor quadrant relative to its own baseline.- Claim 13 and Section 9.2 re-anchored on verified studies of status- and resource-scaled obligation (Trautmann et al. 2023; Berman et al. 2020; Fiddick and Cummins 2007), removing a misattributed anchor.- The quadrant ordering restated as the exact sign statement the functional supports; the identifiability limit stated (flow data identifies only the combined handicap of power and steepness).- Iso-moral-rotation and power-bar figures embedded; stale pre-v16 formulas removed from two figures. New in v18 (2026-07-02):- The evaluation denominator and window as declared parameters of every evaluation (new Section 3.8).- The aggregate position that carries a multi-level agent onto the quadrant plane (new Section 3.4.1).- Composition of the power handicap with the weighted total, a default steepness anchor, and a band protocol for uncalibrated deployments (Section 3.7).- Booking rules for counterfactuals, diverted resources, withdrawal, and revocable access (Sections 6.3 and 6.17 through 6.19).- The agent taxonomy extended to statutes and programs, with attribution rules for compelled and joint conduct (Key Definitions, Section 5.5).- The Field Scoring Protocol with an end-to-end worked example (Section 7.5, Appendix B).- Cognitive sovereignty defined and policy-and-institutional scoring canonized (Sections 10.7, 10.8).- The empirical program split into falsification tests and calibration measurements (Section 9.4).- The moral-work functional is unchanged from v16; all v16 scores, rankings, and archetype classifications carry over. Figure orientation corrected in v18.1 (2026-07-02): every rendering of transmutation space draws amplification on the horizontal axis and filtering on the vertical axis. No theory, claim, or numeric change. New in v19 (2026-07-06):- The accounting stated as normative scoring with declared parameters; Moral Debt, extraction, and the Extractor quadrant framed as ledger positions under declared parameters rather than verdicts on an agent's worth (Section 2.1).- Maslow's hierarchy demoted from foundation to the default filling of a declared need-taxonomy slot; Max-Neef and Doyal-Gough substitute without changing the functional; Claim 4 stated as conditional on the declared taxonomy (Sections 3.1, 11.1).- Sections 8 and 9 restructured for falsifiability: a corrected and extended comparison with the capability approach, prioritarianism, and need theories; a claims-by-type table (Section 8.3); a preliminary first-party section (Section 9.5); and an adverse-evidence subsection covering intent-sensitivity, contested ordering, hedonic adaptation, moral luck, distributive justice, and social measurement (Section 9.3.1).- A causal-attribution graph added to the field scoring protocol, formalizing the marginal-change rule with counterfactual apportionment (Appendix B.7).- A source-linked institutional worked example, the British Columbia hosting of the 2026 FIFA World Cup, added (Appendix B.8).- The moral-work functional is unchanged from v16 through v18.1; all scores, rankings, and archetype classifications carry over. Version history and the provenance of retained constructs are recorded in Appendix C of the paper.