Harlequin: Reputation-Weighted Consensus and Governance for a Voluntary Order

Abstract

Harlequin is a blockchain protocol in which the right to take part in consensus,governance and adjudication comes solely from reputation earned by verifiable acts— never from capital (proof of stake) or expended computation (proof of work).Reputation is a four-dimensional quantity ("the four suits"), computeddeterministically from a public evidence record by a damped trust-propagationfunction, aggregated conservatively (a strong dimension cannot buy authority in aweak one), and subject to time decay so that standing must be continuallyre-earned. Block authorship and committee/jury membership are assigned byreputation-weighted cryptographic sortition; finality is provided by aByzantine-safe gadget over signed votes; disputes are judged by sortitioned jurieswith interest-exclusion, and the only enforced consequence is reputational — theprotocol applies no coercive force. We give the system model, the consensus and justice mechanisms, and a securityanalysis against a state-level adversary whose goal is capture, censorship orde-anonymization rather than direct theft. Two results are emphasized for theirhonesty. First, steady-state Sybil resistance is strong: a Sybil farm withoutearned evidence obtains about 0% of consensus power (17/17 adversarial tests).Second, the cold-start window is not unconditionally safe: a competent adversarypresent at genesis can capture the bootstrap; we show the security of that windowis a race between honest onboarding and adversary mass — bounded, not eliminated,by non-operator personhood verification, an automatic ceiling-halt and theonboarding rate, with the residual risk declared. We report an implementation inRust (dependency-free cores cross-validated against FRAME pallets) and areproducible validation record spanning unit tests and multi-node hardware runs.