Boundary Information Theory (BIT): Four Ontological Postulates, Working Axioms and Sufficient State for Prediction

This record presents a conceptual proposal for Boundary Information Theory (BIT), bringing together four ontological postulates, twenty working axioms and an illustration of sufficient state for prediction. The historical postulates retain their original numbering 0 through 3 and their provenance in the author’s text of 8 September 2026. The Foundation v0.2 distinguishes intrinsic reality, interaction roles, human mathematical representation, purposeful configuration and reasoning across levels. The English and Vietnamese proposal papers are publication edition v0.2. They add a sufficient-state subsection to the earlier v0.1 articles. A bilingual technical companion uses a three-node conductive thermal model to show that equal energy and aggregate temperature need not imply equal boundary flow, and that exact aggregate-temperature prediction need not preserve entropy production. These are conditional consequences of a standard model, used to clarify the information that a representation must retain for selected outputs and admissible inputs. The source archive preserves the original four-postulate document, Foundation versions v0.1 and v0.2, editable publication papers and the Vietnamese technical note BIT-TINH-REGION-001 v0.1. The note’s original machine package was not available in the assembled source set. The principal formulas and values were recalculated, while its nine-check execution and full-trajectory error figures remain source-reported and were not fully replayed for this edition. This is an independent conceptual research report. Logical independence, completeness, novelty and empirical adequacy of the overall BIT framework remain open. The record does not establish a new physical field, quantum-entanglement mechanism or thermodynamic arrow-of-time theory. It introduces no additional research result into the separately scoped, closed BIT Block 6 Stage 1 materials. Ideas and views are by Bùi Quang Trịnh; ChatGPT assisted with organization, source comparison, editing and translation.