CHINA'S SOCIAL CREDIT SYSTEM READ THROUGH THE PHILOSOPHY OF VIRTUES: The Two Handles and the Honour Roll — What Confucius, Han Feizi, Holmström–Milgrom, Bénabou–Tirole, Wintrobe and Tirole's Digital Dystopia Already Hold, and What Remains

Abstract China’s social credit system is usually evaluated as a single object, either as a surveillance scorecard or as a misunderstood regulatory database. Neither reading survives contact with the system as it now stands: there is no unified personal score, most local citizen-scoring pilots are reported to have lapsed, and the system’s documented mass consists of corporate compliance records, court-enforcement blacklists and, since April 2026, a national credit-repair regime. This article evaluates the system component by component, under the programme’s act–state separation, and asks three questions: whether the system serves the freedom of the people who live under it or the power that administers it; how it affects the virtues; and how it affects Chinese society and welfare. Priority is conceded before the argument: for the claim that governing by rewards and punishments produces compliance without shame, to Confucius (Analects 2.3); for the design of the two handles, to Han Feizi; for crowding-out, to Titmuss, Deci and Frey; for the loss of signal value in rewarded prosociality, to Bénabou and Tirole (2006); for diverted effort under partial measurement, to Holmström and Milgrom (1991); for the ruler’s ignorance under repression, to Wintrobe (1998) and Kuran (1995); and for bundling political attitudes into social scores, to Tirole (2021). What remains is structural. The system’s components separate into three classes by a property derived from the programme’s Decoy Placement Rule — whether the substance behind what a component inspects can be taken or directly measured by the scorer — and the classes carry opposite verdicts. Enforcement of owed acts the state can take, measure or compel (debts, taxes, measured emissions) produces no counterfeit virtue and plausible welfare gains, but it frees no one: compelled payment inverts the act it obtains, and seizure, which inverts nothing, removes the occasion on which the obligation could have been kept freely. A general result follows: every increase in a state’s control over its citizens either inverts their acts or removes the occasions for them, and none increases their freedom; control can change states of affairs, but it cannot produce, hold or return freedom, which passes to no one when it is taken. Scoring of displayable conduct (civic points, honour lists, sincerity of apology) is where counterfeit virtue is predicted to concentrate. Bundling of speech into scores transfers the discipline of the scored to the scorer. Eight results follow, including the Honour Roll Theorem (counterfeits concentrate on the reward side of the ledger, not the punitive side), a Double Blindness result joining Wintrobe’s vertical ignorance to the horizontal evidence-displacement found experimentally by Genevsky (2025), and an Occasion Trap showing that joint punishment disables the only unforgeable verifier of reform, which explains why the 2026 repair measures are best read as lifting a sanction rather than certifying a person. Four tests are registered; one uses only public statistics. The sections on Chinese society and welfare are a speculative extension into China studies and empirical political science, fields with their own evidential standards; no specialist in those fields has reviewed this article, and nothing else in the programme may depend on those sections. The article contains no measurement of any human being. Keywords: social credit system; China; Confucius; Han Feizi; Legalism; Holmström–Milgrom; multitask; Bénabou–Tirole; motivation crowding; Tirole; Digital Dystopia; Wintrobe; dictator’s dilemma; Kuran; preference falsification; judgment defaulters; credit repair; redlist; blacklist; sourcehood; occupation; counterfeit virtue; Philosophy of Virtues Highlights · The social credit system is not one object; evaluated by component, its parts carry opposite verdicts on freedom, virtue and welfare. · Enforcement of owed, verifiable acts (debts, taxes, measured emissions) forges nothing but frees no one: compelled payment inverts the act, and seizure removes the occasion for fidelity. · No increase in state control over citizens increases their freedom; freedom taken by a power is not transferred to it but destroyed (Control Theorem, C6). · Counterfeit virtue is predicted to concentrate where the system rewards displays — honour lists and civic points — not where it punishes defaults. · Scores displace acts as evidence of character, for the state (Wintrobe) and between citizens (Genevsky 2025): the Double Blindness result. · Joint punishment removes the occasions on which reform could be shown; the 2026 repair rules rightly key repair to the owed act and elapsed time, not to demonstrated character. Relevance If the article is right, the question “is the social credit system good or bad?” has no answer, and the useful question is which components to keep, which to abolish, and which to rename. The article’s answer is specific: confine enforcement of owed obligations to what is owed, prefer seizure to pressure, and treat it as the least harmful component rather than as a gain for freedom, abolish the scoring of displayable virtue, separate listing from review, and stop calling the enforcement of debts a judgement on character. The same analysis applies without change to Western private scoring, platform reputation and credit-bureau practices wherever they meet the same conditions; the article takes no position on any government, and its criterion is indifferent to which state or firm administers a score. If it is wrong, the programme’s central results on capture and forgery have no purchase on the largest scoring system ever built, and that should be known before those results are used in public argument.