Toward the effective 2-topos
Steve Awodey, Jacopo Emmenegger
TL;DR
This work proposes a concrete path to a realizability-inspired $(2,1)$-topos $\mathcal{E}\!f\!f_2$ whose $0$-types recover the effective topos $\mathcal{E}\!f\!f$, addressing limitations of naive higher-topos constructions in constructive settings. It analyzes $\mathcal{E}\!f\!f$ as the 1-exact completion of the partitioned-assemblies category $\mathcal{P}$ and introduces coherent presheaves $\mathsf{Coh}$, showing $\mathsf{Coh}$ realizes $\mathcal{E}\!f\!f$; then it builds a presheaf-of-groupoids framework and defines coherent groupoids via pseudo-compact diagonals. The main result proves that every coherent fibrant $0$-type $\mathbb{G}$ is equivalent to a discrete coherent object, yielding $\mathsf{Coh}\mathsf{Gpd}(\widehat{\mathcal{P}})_0 \simeq \mathsf{Coh}(\widehat{\mathcal{P}}) \simeq \mathcal{E}\!f\!f$, thereby realizing a candidate $(2,1)$-realizability topos for higher-type foundations. This establishes a pathway toward univalence and higher-type interpretations of realizability in a robust, internal-logical setting with potential practical impact for foundations of higher-type computation and logic.
Abstract
A candidate for the effective 2-topos is proposed and shown to include the effective 1-topos as its subcategory of 0-types.
