Non-Transferability of Justification Across Success Types: Why Succeeding in One Way Does Not Show Success in Another

This conceptual paper introduces Non-Transferability of Justification Across Success Types, a distinction among accuracy, usefulness, authorization, procedural regularity, and justice. Success under one of these standards does not, by the label alone, establish success under another whose conditions differ. The paper asks what kind of success the evidence establishes and what kind the conclusion requires. Connections between them can be justified through additional premises; the distinction does not make them unrelated. It also separates following a procedure from justifying that procedure’s design and suitability. The discussion connects the distinction to Habermas’s account of validity claims and examines the 2014 PNAS Editorial Expression of Concern about Facebook’s emotional contagion study. The statement provides a positive example of keeping editorial acceptability, legal obligation, institutional jurisdiction, and informed-consent concerns separate. It does not find a rule violation or assess the study’s accuracy. Part of the Standards and Answerability series within On Human Modeling, the paper asks which kind of success is allowed to settle a dispute and what supports moving from one assessment to another. Its framework can inform human–AI research on evaluating accuracy, usefulness, authorized use, procedural soundness, and fairness without treating success in one area as sufficient support for the others.