Institutions as Distributed Computations: A Unified Framework Integrating Extended Phenotype Theory, Evolutionary Game Theory, and Computational Complexity

Why do some legal institutions resist reform for decades while others adapt within a single legislative cycle? Existing accounts draw on veto player theory, path dependence, and interest group coalitions. None supplies a mechanism that simultaneously connects individual cognition, collective behavior, and formally specified structural constraints on feasible reform trajectories. This paper proposes such a mechanism through the integration of three research traditions: Extended Phenotype Theory (EPT), Evolutionary Game Theory (EGT), and computational complexity theory. The integration is motivated by a convergence finding: Hamilton et al. (2026), working independently from the EPT-EGT research programme, argue that human societies are distributed computational systems whose organizational forms are shaped by the resource costs of collective problem-solving. Their framework, developed at the intersection of theoretical computer science and social science, maps social coordination phenomena onto formal problems from Boolean circuit theory, distributed computing, approximation complexity, and Byzantine Fault Tolerance. This paper argues that the Constitutional Lock-in Index (CLI), Institutional Hysteresis Rate (IHR), and Institutional Evolvability Index (IEI), developed within the EPT-EGT programme, are operationalizations of the circuit-theoretic concepts Hamilton et al. identify as central to institutional analysis. The paper develops four main contributions. First, a formal mapping between CLI/IHR/IEI values and circuit-theoretic parameters (circuit depth, fault-propagation cost, non-uniform reconfiguration capacity) is proposed and its empirical implications derived. Second, Heteronomous Bayesian Updating (HBU) is reinterpreted as an advice-string mechanism in the sense of computational complexity theory, with specific implications for the timing of reform windows. Third, Zombie Law Equilibria are formally characterized as distributed systems that have exceeded their Byzantine Fault Tolerance threshold, which entails that they cannot be resolved by any internally generated deterministic protocol. Fourth, a research agenda for empirical testing of the synthesis is specified across three methodological platforms: the existing sixty-case international dataset, a proposed natural-experiment design exploiting disruptions to legal professional socialization, and a planned extension of the 816-agent computational model. The paper argues that the EPT-EGT-computation synthesis is more explanatorily powerful than any of its three components separately, because each component supplies what the others lack: EPT supplies the replication mechanism, EGT supplies the selection dynamics, and computational complexity supplies formal constraints on feasible trajectories that neither evolutionary framework can specify on its own.