Independent Researcher — Guardian Research

Defining the comparability layer for independently governed systems.

Two systems can each be valid within their own governance boundaries and still produce judgments that cannot be meaningfully compared.

Comparability before coordination.

My work examines what must be established before judgments from independently governed systems can be meaningfully compared, including whether a valid comparison basis exists at all.

Key questions include:

• Did the systems evaluate the same governed object?
• Did they operate from compatible time states?
• Did they apply compatible definitions and evidence thresholds?
• Were the governing sources and authority conditions preserved?
• Do different outputs represent a genuine conflict, or was comparability never established?

This work does not determine whether a judgment is correct, fair, or authorized for reliance or execution. It determines whether the basis required to admit judgments to comparison has been established. Comparison does not itself authorize reliance, coordination, or execution.

An open-source research project formalizing the conditions required before judgments from independently governed systems can be meaningfully compared or coordinated.

Current areas of investigation include:

• object identity;
• time-state alignment;
• governing definitions;
• evidence thresholds;
• authority and provenance;
• judgment transferability;
• comparison-basis failure; and
• formal incomparability.

Decifact v0.1 is a public research implementation for determining whether independently governed judgments are admitted to comparison.

When comparison is admitted, it distinguishes:

• EQUIVALENT
• NON_EQUIVALENT
• FORMALLY_INCOMPARABLE

When the required comparison basis has not been established, the result is:

• NOT ADMITTED TO COMPARISON

The determination preserves the missing or misaligned comparison conditions, including referent, definition, governing basis, and time-state or currentness conditions.

Decifact does not determine reliance authority, execution admissibility, or final coordination policy.

The public research and reference site for Guardian Research, Guardian, and Decifact.

• Guardian V0.2 — Decision Equivalence
• Guardian V0.3-mini — Acceptance Minimal Invariant
• Guardian V0.3 — Acceptance Architecture and Sovereign Discontinuity Rights
• Guardian V0.4 — Governance Memory Integrity (candidate)

→ View the Guardian repository

The research is informed by an internally operated governance environment that has been running under live execution conditions since February 2026.

It provides an empirical setting for examining how decision boundaries, authority separation, phase progression, and governance continuity behave under operational pressure.

Selected structural findings from that environment have informed questions later formalized through Guardian and Decifact.

Evidence anchoring: RFC 3161 timestamping
Audit structure: Append-only and hash-verified

Current AI governance architectures increasingly address identity, access, policy enforcement, execution evidence, and compliance.

A prior question remains:

When independently governed systems each produce a locally valid judgment, what establishes whether those judgments are comparable—and under what conditions may either judgment be carried into another decision context?

Technical connectivity does not establish judgment transferability.

Different outputs do not, by themselves, establish a meaningful conflict.

Without a comparison basis, coordination may begin before the participating judgments have been shown to share a governable relationship.

Establishing comparability is a prerequisite to meaningful comparison; it is not, by itself, permission to rely on or execute a judgment.

• Email: xsa52000@gmail.com
• LinkedIn: De Yu Chou
• Research site: decifact.com