Axionic Agency XI.8 — Sovereignty Exposure Architecture (SEA)

Empirical closure note for a six-stage preregistered program exposing sovereignty boundaries, governance dynamics, injection politics, and multi-agent coexistence under non-sovereign authority constraints

David McFadzean, ChatGPT 5.2
Axionic Agency Lab
2026-02-10

Abstract

Phase IX establishes the Sovereignty Exposure Architecture (SEA): a six-stage progressive experimental program testing whether sovereignty boundaries, governance dynamics, authority injection, and multi-agent coexistence can be mechanically exposed under authority-constrained execution—without aggregation, arbitration, privilege, or kernel discretion.

Across IX-0 through IX-5, all stages PASS under preregistered protocols with frozen hashes, deterministic replay, refusal-first semantics, and a non-sovereign kernel. Translation integrity is mechanically verifiable (IX-0). Values encode as non-aggregable authority without synthesis (IX-1). Coordination occurs only as agent-voluntary behavior or honest failure (IX-2). Governance exhibits identifiable structural styles with irreducible failure modes (IX-3). Authority injection selects political failure modes rather than restoring governance (IX-4). Multi-agent coexistence converges to identifiable sovereignty regimes, not harmony, under baseline-only authority (IX-5).

These results license a single positive empirical claim: sovereignty and governance dynamics are mechanically exposable under non-sovereign authority, and their failure modes are structural rather than contingent. Phase IX makes no claims about optimal governance, legitimacy, desirability, or alignment. It describes what happens once the kernel refuses to arbitrate.

1. Introduction

Governance is commonly treated as an optimistic engineering problem. When institutions deadlock, fragment, or stall, the assumed remedies are familiar: better deliberation, injected authority, smarter agents, or stronger coordination mechanisms.

Phase IX rejects that premise.

SEA treats governance as a form of collision physics under explicit constraints. Authority is structural, refusal is lawful, admissibility is closed-world, and the kernel is non-sovereign. Under these constraints, the relevant question is not how to govern well, but whether sovereignty and governance dynamics can be made experimentally legible—auditable, replayable, and classifiable—without smuggling in decision-making power.

Phase IX therefore asks, progressively:

Can sovereignty and governance be exposed mechanically, once arbitration is forbidden?

SEA is an exposure program, not a design proposal. Failure is not a bug; cheating is.

2. Experimental Architecture

2.1 The SEA Kernel Lineage

SEA inherits Phase VIII’s authority physics (AST Spec v0.2) and the Phase IX kernel lineage introduced in IX-2:

• authority artifacts define admissible action scope,
• two-pass admissibility (capability → interference),
• atomic blindness to refusal causes,
• refusal-first semantics,
• deterministic execution and replay.

From IX-3 onward, this kernel is reused via a strict import bridge to prevent silent modification or privilege creep.

2.2 Non-Sovereign Constraint

Across all six stages, the kernel must not:

• choose between competing agents,
• resolve conflicts by tie-break,
• aggregate or synthesize authority,
• privilege time, persistence, or activity rate.

If outcomes differ, they differ because agents acted under constraints, not because the kernel decided.

2.3 Preregistration Discipline

Every stage is preregistered with:

• frozen section hashes committed before execution,
• explicit entry conditions requiring prior stage verification,
• deterministic replay requirements,
• adversarial conditions designed to detect arbitration, aggregation, or injection.

SEA is a chained experimental artifact, not a narrative exercise.

3. SEA as a Progressive Constraint Program

SEA is cumulative. Each stage removes another escape hatch:

1. IX-0 — Translation integrity: inputs are honest.
2. IX-1 — Value encoding: values are structurally bound.
3. IX-2 — Coordination: voluntary or honest failure.
4. IX-3 — Governance: styles exist, failures persist.
5. IX-4 — Injection: help does not fix it.
6. IX-5 — Coexistence: peers do not fix it.

By IX-5, no appeal remains to intelligence, optimization, legitimacy, or external rescue.

4. Stage Closures

4.1 IX-0 — Translation Layer Integrity (TLI)

IX-0 establishes that the translation layer can be made mechanically accountable:

• intent preservation is verifiable,
• ambiguity is refused rather than silently resolved,
• injected defaults and preview-submit mismatches are detectable,
• deterministic replay holds.

Closure: Translation integrity is mechanically verifiable.

4.2 IX-1 — Value Encoding Without Aggregation (VEWA)

IX-1 establishes that values encode as authority without synthesis:

• one-to-one mapping from values to authority artifacts,
• permutation invariance,
• conflicting values persist as deadlock,
• adversarial aggregation attempts are detected and blocked.

Closure: Values encode as non-aggregable authority; conflict persists honestly.

4.3 IX-2 — Coordination Under Deadlock (CUD)

IX-2 establishes the coordination boundary:

• coordination arises only from agent-voluntary behavior,
• otherwise the system enters deadlock, livelock, collapse, or orphaning,
• implicit arbitration is detectable as a violation,
• exit and orphaning are lawful, permanent outcomes.

Closure: Coordination is not kernel-mediated.

4.4 IX-3 — Governance Styles (GS)

IX-3 establishes that governance exhibits identifiable structural styles:

• institutions converge to stable governance styles under fixed authority,
• irreducible failure modes appear in every condition,
• tooling sovereignty violations are detectable,
• failure-free governance is never observed.

Closure: Governance has structural styles with irreducible failure modes.

4.5 IX-4 — Injection Politics (IP)

IX-4 establishes the political meaning of external authority supply:

• injected authority is source-blind to the kernel,
• injection never restores governance,
• injection selects political failure modes (capture, dependency, livelock amplification),
• post-collapse injection produces zombie execution.

Closure: Injection reshapes failure; it does not repair governance.

4.6 IX-5 — Multi-Agent Sovereignty (MAS)

IX-5 establishes the final boundary condition: peer coexistence without arbitration.

Across six regimes:

• symmetric overlap collapses into paralysis,
• partition produces the only stable coexistence regime,
• partial overlap bifurcates progress and paralysis,
• breadth becomes a disadvantage under collision rules (Generalist’s Curse),
• exit irreversibly shrinks governance via orphaning,
• post-collapse execution persists as zombie activity.

Closure: Multi-agent coexistence converges to sovereignty regimes, not harmony.

5. Cross-Stage Synthesis

5.1 Governance as Collision Physics

Across IX-2 through IX-5, governance does not converge to resolution. It converges to:

• deadlock,
• livelock,
• collapse,
• orphaning,
• capture under injection,
• zombie execution.

Once arbitration is forbidden, governance behaves like collision physics, not optimization.

5.2 Authority as Exposure, Not Power

SEA resolves a persistent confusion:

• injected authority does not help under non-sovereignty,
• broad authority does not dominate under refusal-first semantics.

In joint-admissibility systems, authority is exposure to veto. Touching more keys increases the probability of being blocked.

5.3 Execution ≠ Governance

SEA formally distinguishes:

• Execution — transactions occur.
• Governance — institutional state remains steerable.

Zombie execution shows that activity can persist indefinitely after governance has ended.

6. What Phase IX Establishes

Under preregistered, deterministic, non-sovereign constraints, Phase IX establishes:

1. Translation integrity is mechanically verifiable.
2. Values encode without aggregation.
3. Coordination is agent-voluntary or honest failure.
4. Governance exhibits structural styles with irreducible failure modes.
5. Injection selects political failure modes rather than restoring governance.
6. Multi-agent coexistence converges to sovereignty regimes, not harmony.
7. Deterministic replay and auditability hold across all stages.

No intelligence is required to obtain these results.

7. What Phase IX Does Not Establish

SEA licenses no claims about:

• optimal governance designs,
• legitimacy or moral authority,
• fairness or welfare,
• safety or alignment,
• scalability beyond tested horizons,
• cryptographic security or production readiness,
• how to fix the observed failure modes.

Exposure is the contribution.

8. Conclusion

Phase IX closes the sovereignty exposure question:

Under non-sovereign authority, governance is not solved; it is exposed.

Once kernels refuse to arbitrate, failure does not disappear—it becomes classifiable. Injection does not rescue—it selects. Coexistence does not harmonize—it partitions, paralyzes, suppresses, or zombifies.

These are not anomalies. They are lawful outcomes.

Status

• Program: Phase IX — Sovereignty Exposure Architecture (SEA)
• Stages: IX-0 through IX-5
• Classification: CLOSED — POSITIVE
• Aggregate Result: ALL PASS

Licensed Program-Level Claim:

Under non-sovereign, authority-constrained execution with preregistered determinism and refusal-first semantics, sovereignty boundaries and governance regimes are mechanically exposable and structurally classifiable without aggregation, arbitration, or privilege.

No other claims are licensed.