When Frontiers Do Not Convert: Contestability, Correction, and Hysteresis in Socio-Technical Change

Evolutionary and frontier-based accounts of technological change ask how innovation expands the space of feasible outcomes and how institutions select among them. For infrastructure-like technologies this paper places a prior question at the centre: under what conditions is frontier technological capacity converted into socially distributed generative capability? Building on, but departing from, evolutionary trajectory theory, utility-frontier approaches, and the recent literature on innovation directionality, the paper develops a Socio-Technical Trajectory Space distinguishing frontier capacity, contestability of deployment assets, correction capacity, generative capability, and welfare. The central claim is that frontier capacity does not imply generative capability: advance becomes socially generative only when deployment is sufficiently contestable and correctable, formally when effective corrective openness exceeds a threshold. Contestability is measured as the traffic-weighted openness of strategic assets, connected here to the political economy of platform infrastructure and to the contestability standard at the centre of recent digital-market regulation. The paper's unifying result is a theorem, not a taxonomy: a single scalar, the enclosure bias, is shown by a lattice-theoretic selection argument to generate three distinct barriers to conversion-a static threshold, a measurement distortion, and a hysteretic exit delay-as three projections of one monotone, memory-bearing object, yielding a falsifiable conversion-reversal threshold.