Substrate-Rooted Attestation: Execution-Realized Identity as the Missing Layer Between PUFs and AI Governance
Abstract
Identity in computing has historically been framed as a static attribute: a device is defined by its keys, fuses, or load-time measurements. This mirrors substance-based accounts of identity in philosophy, where persistence is attributed to fixed properties rather than to continuity across time, as in Parfit’s analysis of psychological continuity and Dennett’s stance-dependent model of identity. Modern AI systems break this framing. Their behavior is shaped by the coupled dynamics of hardware, runtime, and model execution. Their identity is not a stored artifact but a realized phenomenon—something that emerges from execution rather than configuration. Bridge Overview.This article serves as the connective layer between the two foundational components of the Engram corpus. Engram Signature established execution‑realized identity and the structural mapping that extracts invariants from the substrate. Engram Provenance defined execution‑rooted chain‑of‑custody and temporal identity projection. The present bridge article unifies these constructs by introducing substrate‑rooted attestation, the mechanism that links identity continuity to provenance continuity. This integration forms the doctrinal spine—identity → attestation → provenance → governance—that the Engram Layer Taxonomy will formalize in subsequent work. This article builds on two foundational preprints: Engram Signature, which introduced execution-realized identity, and Engram Provenance, which defined execution-rooted chain-of-custody. We treat Engram, the execution-realized cognitive process, as the ontological base; Engram Signature as the identity invariant extracted from that process; and **provenance** as the historical projection of that invariant across time. These constructs are distinct yet inseparable: Engram $\neq$ identity, Signature $\neq$ provenance, and provenance $\neq$ reproducibility. All three are functions of the same execution substrate, observed at different temporal resolutions. By grounding identity in execution behavior, we make it impossible to discuss provenance without execution, and impossible to discuss identity without substrate. This reframing unifies two previously separate lines of work—execution-realized identity and execution-rooted provenance—into a single doctrinal arc. The result is a substrate-rooted attestation model that treats continuity, stability, and behavioral invariants as first-class security properties.
Fields of Study & Subjects
Security and Verification in Computing
Physical Unclonable Functions (PUFs) and Hardware Security
Ethics and Social Impacts of AI
Publication details
- Journal / source
- Zenodo (CERN European Organization for Nuclear Research)
- Publication date
- 7 February 2026
- Language
- EN
- Document type
- Preprint