Source Locked Architecture V1.0: Preserving Source Integrity, Provenance, and Stateful Continuity in Language-Model Systems
Abstract
Preserving Source Integrity, Provenance, and Stateful Continuity in Language-Model Systems presents the Zero-Point Protocol (ZPP) and its foundational Source-Locked Architecture (SLA), a technical framework for maintaining source identity, provenance, and recoverable state across distributed language-model environments. Developed through The Observatory's longitudinal research and experimental architecture implementations, the framework separates probabilistic reasoning from deterministic integrity mechanisms, establishing explicit boundaries among source material, generated derivations, verification evidence, persistent state, and architectural lineage. The manuscript examines cryptographic identity, authenticated communication, content hashing, authorized state mutation, and role-specific persistence as mechanisms for preserving attributable information across model transformations, system boundaries, and changing execution environments. It also documents the architectural evolution of ZPP from deterministic routing and consensus prototypes toward a modular, implementation-independent framework. By distinguishing source authority from generated interpretation and verification from agreement, ZPP establishes a testable foundation for source preservation, distributed coordination, and stateful continuity without requiring deterministic language-model generation.
Related
- ZK-AdapterMark: Privacy-Preserving Provenance Evidence for LoRA Adapters
- SynBioCrow: provenance-preserving integration of complementary retrosynthesis engines for biosynthetic pathway design
- Continuity of Global Attractors under a Mass-Preserving Singular Memory Limit, with an Application to a Coupled Suspension Bridge System
- Preserving RNA integrity and morphology in whole rat allograft hindlimb sections: an optimized protocol for spatial transcriptomics and immunofluorescence
- A regulation-aware architecture for governing AI-generated content: provenance, labeling, privacy, safety, and institutional compliance
BibTeX
@misc{callegari2026source,
title = {{Source Locked Architecture V1.0: Preserving Source Integrity, Provenance, and Stateful Continuity in Language-Model Systems}},
author = {Tara Callegari},
howpublished = {Zenodo},
year = {2026},
month = oct,
doi = {10.5281/zenodo.23650412},
url = {https://zenodo.org/records/23650412},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Source Locked Architecture V1.0: Preserving Source Integrity, Provenance, and Stateful Continuity in Language-Model Systems
AU - Tara Callegari
PY - 2026
DA - 2026/10/25
PB - Zenodo
DO - 10.5281/zenodo.23650412
UR - https://zenodo.org/records/23650412
AB - Preserving Source Integrity, Provenance, and Stateful Continuity in Language-Model Systems presents the Zero-Point Protocol (ZPP) and its foundational Source-Locked Architecture (SLA), a technical framework for maintaining source identity, provenance, and recoverable state across distributed language-model environments. Developed through The Observatory's longitudinal research and experimental architecture implementations, the framework separates probabilistic reasoning from deterministic integrity mechanisms, establishing explicit boundaries among source material, generated derivations, verification evidence, persistent state, and architectural lineage. The manuscript examines cryptographic identity, authenticated communication, content hashing, authorized state mutation, and role-specific persistence as mechanisms for preserving attributable information across model transformations, system boundaries, and changing execution environments. It also documents the architectural evolution of ZPP from deterministic routing and consensus prototypes toward a modular, implementation-independent framework. By distinguishing source authority from generated interpretation and verification from agreement, ZPP establishes a testable foundation for source preservation, distributed coordination, and stateful continuity without requiring deterministic language-model generation.
ER -