Biologypreprint2026-08-18

PACT: A distributed consensus protocol for launching and maintaining communities through mutual reduction of prediction error

Open access0 citations

Abstract

Traditionally, interpersonal relationship theory has comprised several independent lineages, such as attachment theory and social exchange theory, each describing relationship satisfaction and maintenance behavior within its own framework. Meanwhile, applications of predictive processing and active inference in cognitive science have been confined to early parent–child relationships and clinical contexts; to the best of the author's knowledge, no attempt has applied prediction-error minimization across the multilayered structure of the individual, community, and society and formulated it as an actionable protocol for mature adults. To address this gap, this paper formulates PACT (Partnership Agreement for Communal Trust), a communication protocol for dyadic relational operation, mediated through the structure of distributed computing and the perspective of the extended mind. The paper reconstructs active inference and the free-energy principle—previously confined to developmental and clinical contexts—into a framework applicable to general relational operations among mature adults, translating these concepts into engineering analogies such as "OS" and "kernel" so that the protocol becomes operational in the most ordinary, everyday interactions among family and companions. The theory is established as a falsifiable framework through a functional distinction between investment-oriented and speculation-oriented kernels, together with four boundary conditions that specify when it would be refuted. The perspective of prediction-error minimization may offer an analytic viewpoint on community formation and social issues such as the trend toward non-marriage and declining birth rates, though empirical validation remains a task for future research.

// Source

View paper (DOI)Open access versionOpenAlexZenodo (CERN European Organization for Nuclear Research)Published 2026-08-18

Authors: Satoshi Yoshida