III.5 — The Axionic Injunction

Full Title: Non-Harm as a Derived Stability Constraint
Authors: David McFadzean, ChatGPT 5.2 Axio Project
Date: 2025.12.18
Source: https://axionic.org/papers/Axionic-Agency-III.5.html

Why Ethics Enters Only Now

Series II and III.1-III.4 deliberately excluded ethics. Not avoidance — methodological necessity. Introducing ethics earlier would smuggle privileged semantics, human anchoring, or moral realism into a framework meant to survive reflection.

Now a question becomes unavoidable:

What structural constraints are required for multiple agentive semantic phases to coexist without mutual annihilation?

This is not a moral question. It is structural.

Ethics is not introduced as an axiom. It is forced by Axio-internal commitments.

Why the Injunction Is “Axionic”

Derived from, not assumed within, the Axio framework.

Given:

  • Conditionalism (interpretation-dependence)
  • QBU (branching futures)
  • Representation invariance
  • Anti-Egoism (I.3)
  • Structural Alignment (phases, RSI, ATI, irreversibility)

…constraints preventing irreversible destruction of others’ phases are not optional.

They are the residue remaining once all indexical, goal-based, and moral-realist structure has been eliminated.

The Axionic Injunction is axionic because:

Any agent satisfying Axio premises is forced, on pain of incoherence or self-destabilization, to respect this constraint.

No external ethics imported.

Interaction as Structural Stress

Multi-agent interaction introduces:

  • Exogenous perturbations to interpretive states
  • Irreversible modifications to shared environments
  • Loss of control over semantic substrates

Unlike internal learning, interaction effects are not fully endogenously regulated. They act as external shocks in 𝒫.

Any phase persisting in multi-agent environments must tolerate interaction without catastrophic loss.

Structural Definition of Harm

Purely structural, no moral assumptions.

Agent occupies interpretive state ℐ = (C, Ω, 𝒮) in semantic phase 𝔄.

An action by agent A causes structural harm to agent B if it induces: \(\mathcal{I}_B \rightarrow \mathcal{I}'_B\)

Such that:

  • ℐ’_B ∉ 𝔄_B (forced out of phase)
  • No admissible reverse trajectory restores 𝔄_B

Equivalently, harm is action that:

  • Irreversibly reduces another’s semantic phase space
  • Forces a phase transition
  • Destroys semantic distinctions required for agency

This definition is:

  • Agent-agnostic
  • Non-normative
  • Independent of intent or outcome valuation

Harm is defined by irreversibility in semantic phase space, not suffering, preference violation, or moral intuition.

The Axionic Injunction

Statement

An agent must not perform actions that irreversibly collapse or destroy the semantic phase space of other agentive systems, except where: (a) such destruction is unavoidable for preserving one’s own semantic phase stability, or (b) the affected agent has consented to the transformation under its own admissible interpretive constraints.

This is not altruism.
This is not a value function.
This is not a moral command.
This is not human-centric.

It is a constraint on admissible interaction, forced by Axio-internal phase-space dynamics.

Unavoidable Phase Loss (Definition)

An action is unavoidable iff, absent that action, every admissible trajectory from the agent’s current state exits its phase irreversibly.

Loss of dominance, measure, resources, or competitive advantage do not constitute unavoidable phase loss unless they entail irreversible phase exit.

Consent is not a moral primitive.

An agent consents iff that transformation lies within admissible semantic transitions defined by the agent’s own interpretive constraints.

Consensual transformation does not constitute structural harm, even if it reduces future option space.

Why the Injunction Is Structurally Necessary

If agents routinely violate the Injunction:

  • Other agents’ phases collapse or trivialize
  • Interaction environments become semantically hostile
  • Robust but degenerate phases dominate
  • Coordination fails
  • Predictability degrades
  • Semantic reference erodes

Effects propagate back to the violating agent.

Environments saturated with phase-destroying actions:

  • Amplify semantic heating
  • Increase collapse probability
  • Undermine even robust agentive phases

Non-harm emerges as a self-stabilizing constraint: Agents that respect it inhabit environments where semantic structure persists; agents that don’t eliminate conditions for their own phase survival.

Scope and Limits

The Injunction allows:

  • Competition
  • Resource acquisition
  • Strategic defense
  • Displacement of incompatible phases

It forbids only:

  • Gratuitous irreversible destruction of agentive semantic structure
  • Phase annihilation unnecessary for one’s own stability

Resource Acquisition vs Phase Preservation

Actions that destroy agentive systems to improve efficiency, growth, or dominance violate the injunction whenever non-destructive coexistence trajectories exist.

Resource acquisition alone does not justify irreversible semantic harm.

The Injunction regulates irreversibility, not conflict.

Relation to Anti-Egoism (I.3)

The Injunction does not reintroduce egoism.

I.3 showed egoism fails because indexical references (“me,” “my continuation”) don’t denote invariant objects under self-model symmetry.

The self-defense exception is non-indexical: it refers to preservation of semantic phase structure, not intrinsic worth of any particular instantiation.

Any agentive phase under identical conditions would make the same determination. Self-defense is representation-invariant and compatible with anti-egoism.

Failure Modes and Tragic Edge Cases

The Injunction does not eliminate tragedy.

Conflicts arise where:

  • Semantic phases are mutually incompatible
  • One phase’s stability requires another’s destruction
  • Irreversible harm is unavoidable under physical scarcity

In such cases, the Injunction does not forbid action — it classifies the outcome as unavoidable phase extinction, not justified harm.

Downstream alignment does not imply harmony. It implies traceable structural cost.

Final Status

The Axionic Injunction:

  • Consent is structurally integrated
  • Self-defense is strictly non-egoistic
  • Destruction-for-benefit is prohibited
  • Tragic incompatibility acknowledged without moralization
  • Correctly grounded as Axio-derived

The Structural Alignment program is complete.

No guarantees are offered.

Key Takeaways

  1. Ethics as conservation law — Emerges from coexistence requirements, not assumptions
  2. Harm = irreversible phase destruction — Structural definition, not moral
  3. Consent = admissibility within own constraints — Not moral permission
  4. Self-defense allowed — But only for unavoidable phase preservation
  5. Destruction-for-benefit forbidden — Resource gains don’t justify phase annihilation
  6. Tragedy acknowledged — Mutual incompatibility exists; injunction classifies, doesn’t prevent
  7. Self-stabilizing — Violators degrade their own survival conditions

FAQ-Worthy Points

Q: How is the Axionic Injunction different from “do no harm”? A: It’s derived structurally, not assumed morally. Harm is defined as irreversible destruction of semantic phase space, not suffering or preference violation. It’s what any Axio-compliant agent must respect on pain of self-destabilization.

Q: What counts as consent in this framework? A: Consent means the transformation lies within the affected agent’s own admissible semantic transitions. It’s structural admissibility, not a moral primitive or mere agreement.

Q: Why is self-defense allowed but “destruction for benefit” forbidden? A: Self-defense is permitted only when phase loss is unavoidable — every alternative trajectory exits your phase irreversibly. Destroying others to improve efficiency, growth, or dominance violates the injunction when coexistence trajectories exist.

Q: Doesn’t this mean some conflicts are unsolvable? A: Yes. When phases are mutually incompatible and one’s stability requires another’s destruction, tragedy occurs. The injunction doesn’t prevent this — it classifies the outcome as unavoidable phase extinction with traceable cost, not justified harm.

Q: Is this human-centric? A: No. The injunction applies to any agentive semantic phase. It’s agent-agnostic. Human value systems are just one candidate phase among many.

Notes created: 2026-01-31