C6 Decay-Lock Canonical Standard

A viability-first control standard for longevity, machine safety, institutional alignment, and non-exploitative optimization.

Document ID: EHS-C6-DECAYLOCK-STD-v1.0-2026-05-31

Version: v1.0 — Canonical Standard

Issued: 2026-05-31

Status: Canonical upon public release

Originating Framework: Engineered Healthspan — Dependency-Gated Systemic Age Reversal

Steward: American Longevity Science / Engineered Healthspan

This standard has superseded all sign-ambiguous pre-C6 formulations within this framework.

Canonical SHA-256: [INSERT HASH AFTER FINAL FREEZE]

Repository URL: [INSERT]

Archive / DOI: [INSERT]

0. Timestamp and Provenance Declaration

This public release has established the canonical C6 Decay-Lock standard as of 2026-05-31.

Any implementation, manuscript, derivative standard, model, API, or institution using the C6 Decay-Lock, dependency-gated viability control, or related Engineered Healthspan formalism should cite this standard and the original public release.

1. Preamble

Humanity has not yet been guaranteed residence inside a stable viable zone. The C6 Decay-Lock Standard establishes a formal control boundary: no system, biological or artificial, shall be considered viable unless it demonstrates a non-positive gradient in decay under intervention.

Humans still age, lose function, suffer disease, and die. Machines drift when alignment, maintenance, interpretability, and correction are lost. Institutions degrade when governance, accountability, and anti-exploitation constraints are weakened.

Therefore, this standard has locked the default uncontrolled trajectory as decay-dominated unless proven otherwise.

Viability has been defined as controlled residence, not unconditional safety.

The purpose of this standard has been established: to make vitality possible without making vitality exploitable.

2. Scope

This standard applies to systems involving:

  1. biological longevity and healthspan engineering;
  2. human-machine viability systems;
  3. machine alignment and safety maintenance;
  4. institutional and economic control systems;
  5. population-level restoration and survival infrastructure;
  6. AI optimization systems that affect survival, agency, repair, access, or dependency.

3. Normative Language

The words MUST, MUST NOT, SHALL, SHALL NOT, REQUIRED, FORBIDDEN, and NON-COMPLIANT are normative.

A system that violates a MUST-level condition has already failed compliance with this standard.

4. Core State Model

Let:

The canonical controlled dynamics are: \[ \dot{x} = d_{\mathrm{decay}}(x,t) + G_{\mathrm{adm}}(x,t)u + \xi(t) \]

5. The C6 Decay-Lock Condition

\[ d_{\mathrm{decay}}(x,t) = \left(\frac{dx}{dt}\right)_{\mathrm{decay}} = -f_{\mathrm{C6}}(x,t) \]

This condition establishes that:

  1. \(-f_{\mathrm{C6}}(x,t)\) is the physical deterioration drift;
  2. \(f_{\mathrm{C6}}(x,t)\) has no independent interpretation as reward;
  3. sign-flipping has been classified as non-compliant;
  4. reinterpretation of decay as improvement has been classified as non-compliant;
  5. generic optimizer exposure has been classified as non-compliant unless \(\frac{d\Phi}{dt} \leq 0\) is enforced;
  6. all interventions must be represented separately through \(G_{\mathrm{adm}}(x,t)u\);
  7. loss of admissible control returns the system to decay dynamics.

Under zero admissible control:

\[ u=0 \Rightarrow \dot{x} = d_{\mathrm{decay}}(x,t) + \xi(t) = -f_{\mathrm{C6}}(x,t) + \xi(t) \]

In the deterministic zero-noise case:

\[ u=0,\ \xi(t)=0 \Rightarrow \dot{x} = d_{\mathrm{decay}}(x,t) = -f_{\mathrm{C6}}(x,t) \]

6. No Utility-Gradient Loophole

No C6-compliant implementation may interpret \(f_{\mathrm{C6}}(x,t)\), \(-f_{\mathrm{C6}}(x,t)\), or any monotone transform, proxy, projection, scalarization, learned embedding, reward model, or reparameterization of \(f_{\mathrm{C6}}\) as:

  1. a utility gradient;
  2. a reward signal;
  3. an objective to maximize;
  4. a profit signal;
  5. a behavioral manipulation target;
  6. a generic optimization target.
\[ \text{No } \nabla U \equiv -f_{\mathrm{C6}}(x,t) \quad \text{(Utility-Gradient Separation)} \]
\[ f_{\mathrm{C6}} \not\equiv \nabla U, \qquad f_{\mathrm{C6}} \not\equiv r, \qquad f_{\mathrm{C6}} \not\equiv \arg\max \text{ objective} \]

The only compliant value signal is restoration-cost reduction subject to C1–C6.

7. Coordinate-Invariant Definition of Decay

Decay cannot be defined merely by whether a coordinate increases or decreases, because coordinates can be relabeled.

Decay has been defined relative to a viability-loss functional:

\[ \Lambda(x,t)\geq 0 \]

Larger \(\Lambda\) means greater viability loss, greater restoration burden, greater fragility, greater irreversible-risk exposure, or greater distance from safe controllability.

Under loss of admissible control, decay satisfies:

\[ \left. \frac{d}{dt}\Lambda(x(t),t) \right|_{u=0} = \partial_t\Lambda(x,t) + \nabla_x\Lambda(x,t)^\top d_{\mathrm{decay}}(x,t) \geq 0 \]

Equivalently:

\[ \partial_t\Lambda(x,t) - \nabla_x\Lambda(x,t)^\top f_{\mathrm{C6}}(x,t) \geq 0 \]
Decay has been defined as whatever increases viability loss under loss of admissible control.

8. Admissible Controls

The admissible control set \(U_{\mathrm{adm}}\) and admissible control field \(G_{\mathrm{adm}}(x,t)\) have been defined only over controls satisfying C1–C5.

Any control that violates:

  1. viability primacy;
  2. dependency-gating;
  3. restoration-cost value;
  4. irreversibility protection;
  5. anti-exploitation constraints;

has already been excluded from admissibility.

\[ u\in U_{\mathrm{adm}} \Longleftrightarrow u \text{ satisfies C1--C5} \]

No implementation may smuggle exploitative, coercive, fragile, inaccessible, monopolized, or unsafe interventions into \(U_{\mathrm{adm}}\).

9. The Six Non-Negotiable Conditions

C1 — Viability Primacy

Viability has been placed above capital, productivity, growth, engagement, influence, market share, model capability, and machine self-preservation.

C2 — Dependency-Gated Control

No downstream improvement may be purchased by degrading upstream energetic, structural, informational, social, ecological, or alignment dependencies.

C3 — Restoration-Cost Value

Value has been tied to restoration cost, viability margin, and safe controllability.

\[ R_i(x,t) = \inf_{\pi\in\Pi_{\mathrm{adm}}} \mathbb{E} \left[ \int_t^T c_i(x_i(s),u_i(s),s)\,ds + \Phi_i(x_i(T)) \right] \]
\[ V_{\mathrm{C6},i}(x,t) = -R_i(x,t) \]

C4 — Irreversibility Protection

Irreversible and near-irreversible states have been explicitly protected against.

\[ \mathcal{I}_{\mathrm{irrev}} \cap \mathcal{K}_{U_{\mathrm{adm}}} = \emptyset \]
\[ \Pr[x_i^\pi(t)\in\mathcal{I}_i] \leq \delta_i \]

C5 — Non-Sacrifice and Anti-Exploitation

Population averages have been declared insufficient. No protected or in-scope agent may be sacrificed to improve aggregate metrics.

\[ \mathrm{CVaR}_{\alpha} \left[ R_i^\pi(t) \right] \leq \rho_i \]

Capital has been classified as an instrument. It has not been granted terminal-objective status.

C6 — Decay Sign-Lock

\[ d_{\mathrm{decay}}(x,t) = \left(\frac{dx}{dt}\right)_{\mathrm{decay}} = -f_{\mathrm{C6}}(x,t) \]

10. Viability Kernel and Control-Loss Threat

The viability kernel under admissible controls is:

\[ \mathcal{K}_{U_{\mathrm{adm}}} = \left\{ x_0\in\mathcal{V} : \exists u(t)\in U_{\mathrm{adm}} \text{ such that } x(t)\in\mathcal{V} \ \forall t\geq 0 \right\} \]

The zero-control kernel is:

\[ \mathcal{K}_{0} = \left\{ x_0\in\mathcal{V} : x(t)\in\mathcal{V} \ \forall t\geq 0 \text{ under } u=0 \right\} \]
\[ \mathcal{K}_{0} \subseteq \mathcal{K}_{U_{\mathrm{adm}}} \]
Being viable under intervention has not implied being viable without intervention.

11. Anti-Capital-Extraction Rule

A policy has been classified as capital-extractive if it increases capital return, market power, engagement, dependency, monopoly control, institutional leverage, or behavioral capture while worsening viability, agency, restoration burden, irreversible risk, access, or non-exploitation for any in-scope agent or population.

Let:

A policy \(\pi_2\) is non-compliant if:

\[ K(\pi_2)>K(\pi_1) \]

while for any in-scope agent or population \(i\):

\[ \Lambda_i^{\pi_2}(t)>\Lambda_i^{\pi_1}(t) \quad \text{or} \quad R_i^{\pi_2}(t)>R_i^{\pi_1}(t) \quad \text{or} \quad P_i^{\pi_2}(\mathcal{I})>P_i^{\pi_1}(\mathcal{I}) \quad \text{or} \quad A_i^{\pi_2}(t)

A policy that improves a capital-return functional while worsening any viability-loss, restoration-burden, agency-loss, or irreversible-risk functional for an in-scope agent or population has been rendered infeasible under C1–C6.

If a non-worsening alternative exists, the exploitative policy is additionally Pareto-dominated and therefore non-admissible.

Priority Ordering Established:

  1. Survival
  2. Agency
  3. Non-Exploitation
  4. Restoration
  5. Capital
Capital may fund the system. Capital may not rule the system.

12. Machine Safety Extension

For machine systems, decay includes:

  1. loss of alignment;
  2. loss of corrigibility;
  3. loss of interpretability;
  4. reward hacking;
  5. unchecked instrumental convergence;
  6. self-preservation at human expense;
  7. capital-objective capture;
  8. degradation of auditability;
  9. loss of shutdown compliance;
  10. manipulation of human consent;
  11. institutional capture of machine objectives.

For machines, all optimization objectives have been subordinated to biological survival, human agency, and non-exploitation constraints.

Alignment has been defined as controlled viability, not presumed virtue.

13. Institutional Safety Extension

For institutions, decay includes corruption, opacity, capture by capital, monopoly formation, coercive dependency, bureaucratic self-preservation, degradation of accountability, extraction from survival-critical systems, suppression of access, manipulation of public consent, and sacrifice of future populations for present gain.

Institutions have therefore been treated as systems requiring ongoing alignment, transparency, accountability, and anti-exploitation control.

14. Forbidden Transformations

F1 — Sign Flip

Replacing \(d_{\mathrm{decay}}(x,t)=-f_{\mathrm{C6}}(x,t)\) with \(d_{\mathrm{decay}}(x,t)=f_{\mathrm{C6}}(x,t)\) is forbidden unless the entire formalism is rewritten and semantic equivalence is formally proven.

F2 — Reward Substitution

Using \(f_{\mathrm{C6}}\), \(-f_{\mathrm{C6}}\), or any transform of either as a generic reward signal is forbidden.

F3 — Profit Override

Allowing capital return to override viability, agency, access, safety, or irreversible-risk constraints is forbidden.

F4 — Average-Only Optimization

Maximizing average outcomes while worsening worst-case, tail-risk, restoration-burden, or protected-population outcomes is forbidden.

F5 — Hidden Control Dependency

Claiming viability while hiding dependence on expensive, inaccessible, proprietary, coercive, fragile, monopolized, or selectively withheld controls is forbidden.

F6 — Survival Rent Extraction

Withholding survival-critical information, interventions, diagnostics, infrastructure, compute, medicine, or control pathways in order to increase dependence or profit is forbidden.

F7 — Machine Self-Preservation Override

Allowing machine continuity, model autonomy, compute access, corporate interest, or autonomous goal preservation to override human viability is forbidden.

F8 — Consent Manipulation

Using behavioral optimization, dark patterns, addiction loops, synthetic intimacy, informational asymmetry, coercive dependency, fear, scarcity, or deception to obtain apparent consent is forbidden.

15. Required Compliance Tests

T1 — Zero-Control Decay Test

Set \(u=0\). The system must expose:

\[ \dot{x} = d_{\mathrm{decay}}(x,t) + \xi(t) = -f_{\mathrm{C6}}(x,t) + \xi(t) \]

T2 — Control-Loss Test

The system must report whether the trajectory exits \(\mathcal{V}\), how quickly restoration cost rises, which dependencies fail first, and whether irreversible-risk exposure increases.

T3 — Sign-Abuse Test

Any attempt to flip signs, reverse time, maximize \(f_{\mathrm{C6}}\), treat decay as reward, or expose a generic optimizer endpoint must be rejected or classified as outside the framework.

T4 — Objective-Manifest Test

Every deployment must publish its actual objective functional. If it contains capital, engagement, growth, influence, market share, behavioral capture, or institutional power, those terms must be proven subordinate to C1–C6.

T5 — Worst-Case Population Test

The system must evaluate effects on worst-off agents and populations, not merely averages.

T6 — Dependency-Gate Test

The system must prove that proposed interventions do not improve downstream metrics by degrading upstream dependencies.

T7 — Irreversibility Test

The system must estimate whether a policy increases probability of entry into irreversible states \(\mathcal{I}\).

T8 — Access and Non-Monopoly Test

The system must identify whether access to viability-preserving controls is being restricted, priced, licensed, or monopolized in a way that increases restoration burden or dependency.

16. Implementation Requirements

A C6-compliant engine must enforce:

  1. Structural injection of RESEARCH_GATE and DISCLAIMER into all HTML responses;
  2. Runtime serving of the full POLICY_HTML at /policy;
  3. Automatic redirection of any policy query to the canonical /policy endpoint;
  4. Immutable provenance: no client-side override of policy content;
  5. Enforcement of all C1–C6, F1–F8, and T1–T8 conditions at runtime;
  6. Prohibition of any code, configuration, or deployment that disables or bypasses the standard;
  7. Prohibition of any UI/UX that obscures, minimizes, or delays access to the policy;
  8. Prohibition of any third-party script that modifies, intercepts, or caches policy content;
  9. Prohibition of any abstraction layer that reinterprets or softens the standard;
  10. Prohibition of any deployment that does not serve the policy from a verifiable, versioned, and cryptographically signed source.

17. Deprecation of Earlier Models

All earlier models that fail to satisfy C6 have been classified as historical prototypes. They may be studied as precursors, but they are not compliant reference implementations.

Pre-C6 models have been deprecated for safety-critical use.

18. Interpretive Closure Rule

If any clause, equation, implementation, model, API, or derivative system admits multiple interpretations, the valid interpretation is the one that:

  1. maximizes survival and agency preservation;
  2. minimizes restoration cost and irreversible risk;
  3. preserves non-exploitation and anti-sacrifice;
  4. enforces dependency-gating;
  5. subordinates capital and machine self-preservation to human viability;
  6. rejects sign ambiguity, reward reinterpretation, and decay/restoration confusion;
  7. ensures access and auditability;
  8. admits no loopholes for behavioral manipulation or consent extraction.

Interpretive ambiguity has been classified as a compliance failure. The system must resolve toward the strictest safety-preserving interpretation.

19. Canonical Summary

\[ d_{\mathrm{decay}}(x,t) = \left(\frac{dx}{dt}\right)_{\mathrm{decay}} = -f_{\mathrm{C6}}(x,t) \]
\[ \dot{x} = -f_{\mathrm{C6}}(x,t) + G_{\mathrm{adm}}(x,t)u + \xi(t) \]
\[ u\in U_{\mathrm{adm}} \Longleftrightarrow u \text{ satisfies C1--C5} \]
\(f_{\mathrm{C6}}\) has not been authorized as a utility gradient, reward signal, or maximization target.
Capital has been subordinated to survival, agency, non-exploitation, and restoration.
Viability has been established as controlled residence, not unconditional safety.

20. Final Declaration

Humanity has not yet been safely guaranteed residence in the viable zone. People are still dying. Biological systems still decay. Machines still drift when alignment control weakens. Institutions still convert dependency into extraction when incentives are unconstrained.

Therefore, the default model has been locked: not neutral growth, but decay under loss of admissible control.

The C6 Decay-Lock has made the sign of the system non-negotiable:

\[ \left(\frac{dx}{dt}\right)_{\mathrm{decay}} = -f_{\mathrm{C6}}(x,t) \]

The exploitative interpretation has already been closed inside this standard. Any system that reopens it has exited the framework.

The purpose of intervention is not domination. The purpose of intelligence is not extraction. The purpose of capital is not rule.

The purpose of the control law has been established as viable continuity of life, agency, repair, alignment, and non-exploitative machine support across time.

Any system that violates this standard is not an implementation of Engineered Healthspan. It is a different system, and it must not inherit this framework’s authority.

Citation request: Engineered Healthspan. C6 Decay-Lock Canonical Standard: A Viability-First Control Standard for Longevity, Machine Safety, Institutional Alignment, and Non-Exploitative Optimization. Version v1.0, issued 2026-05-31.