TA-14 — Controlled Definitions

TA-14 uses controlled language because execution governance cannot be built on loose terminology.

A system cannot enforce what it has not defined. A record cannot become admissible if the conditions of admissibility are unclear. An action cannot be governed if the meaning of “state,” “authority,” “continuity,” “binding,” or “execution” changes from one context to another.

This page establishes the controlled definitions used within the TA-14 Admissible Execution Standard.

These terms are not descriptive preferences. They are structural terms. They define the conditions under which a system, person, institution, control layer, automation layer, artificial intelligence system, maintenance process, environmental process, or operational workflow may rely on recorded truth and proceed toward action.

TA-14 does not treat language as decorative. Language defines the boundary of reliance.

Where terms are undefined, systems drift.

Where systems drift, execution becomes unbound.

Where execution is unbound, action may occur without admissible truth.

TA-14 exists to prevent that condition.

Admissible Execution

Admissible Execution is the condition in which an action is permitted only because the supporting state has been established, preserved, validated, bound, and confirmed as admissible at the moment execution is attempted.

Execution is not admissible merely because a system has data.

Execution is not admissible merely because a person believes the action is reasonable.

Execution is not admissible merely because a recommendation, model output, sensor value, checklist, report, or prior decision exists.

Execution becomes admissible only when the action is bound to a valid admissible state at commit time.

Admissible Execution requires:

1. Origin-captured evidence;

2. Preserved chronological continuity;

3. Valid authority;

4. Defined scope;

5. A binding object connecting state to action;

6. Commit-time enforcement;

7. Outcome recording.

If any required condition is absent, expired, contradicted, reconstructed, or outside its valid scope, execution must not proceed as admissible execution.

Admissible State

Admissible State is the condition of a system, environment, record, person-associated exposure, process, asset, or governed context that may be relied upon for action because it has been established through admissible evidence.

An admissible state is not a guess.

It is not a snapshot.

It is not an assumption.

It is not a model-generated conclusion.

It is not a reconstructed explanation after the fact.

An admissible state is a state whose origin, timing, continuity, scope, and authority are sufficiently preserved to permit reliance.

In TA-14, admissible state must be current enough, continuous enough, bounded enough, and authority-valid enough to support the specific action being considered.

A state may be admissible for one action and inadmissible for another.

A state may be admissible for observation but not for intervention.

A state may be admissible for awareness but not for automated execution.

A state may be admissible for human review but not for non-bypassable control.

Admissibility is not universal. It is bound to use.

Commit Time

Commit Time is the final moment immediately before execution when a proposed action becomes irreversible, materially consequential, externally binding, or system-affecting.

Commit time is not the time when a recommendation was generated.

It is not the time when a report was written.

It is not the time when a sensor reading was captured.

It is not the time when a person approved a plan earlier in the workflow.

Commit time is the moment the system is about to act.

At commit time, TA-14 requires the system to confirm that the supporting state remains admissible.

A prior admissible state cannot be silently reused.

A prior approval cannot substitute for present admissibility.

A cached conclusion cannot substitute for commit-time validation.

A recommendation that was valid earlier may become invalid if continuity is broken, authority changes, evidence expires, scope changes, or drift occurs.

Commit-time admissibility is the central enforcement condition of TA-14.

Execution Boundary

The Execution Boundary is the non-bypassable structural boundary between proposed action and actual action.

It is the point at which a system must determine whether execution is allowed, blocked, or escalated.

The execution boundary is not a dashboard.

It is not a suggestion layer.

It is not a compliance note.

It is not a post-event review.

It is not a risk score.

The execution boundary is where authority to act is either granted or withheld.

In TA-14, the execution boundary must be governed by admissible state. It must not permit action based on incomplete evidence, reconstructed timelines, unbound recommendations, stale records, inferred authority, or model confidence alone.

A valid execution boundary must be capable of refusing execution.

If a system cannot block action when admissibility fails, it does not contain a TA-14 execution boundary.

Continuity

Continuity is the preserved chronological relationship between origin evidence, state, admissibility, authority, binding, execution, and outcome.

Continuity is not the mere existence of stored data.

Continuity is not periodic monitoring.

Continuity is not a collection of disconnected logs.

Continuity requires that the record preserve the sequence of relevant events without unauthorized gaps, substitutions, retrospective reconstruction, or silent alteration.

In TA-14, continuity determines whether a record can support reliance.

A record with unexplained gaps may still be useful for awareness, review, or investigation, but it does not automatically support admissible execution.

Continuity must be preserved before interpretation.

Continuity must be preserved before authority.

Continuity must be preserved before binding.

Continuity must be preserved before execution.

Where continuity is broken, reliance must be limited, paused, blocked, or escalated according to the governing standard.

Origin Capture

Origin Capture is the recording of evidence at or near the source event, condition, measurement, observation, intervention, transition, or environmental state being preserved.

Origin capture is required because reconstructed truth is structurally weaker than preserved truth.

TA-14 gives priority to evidence captured at origin because admissible execution depends on knowing what was true before interpretation, compression, omission, alteration, or institutional preference shaped the record.

Origin capture does not mean every record is perfect.

It means the system distinguishes between what was captured as it occurred and what was inferred later.

That distinction is essential.

A system that cannot distinguish origin-captured evidence from reconstructed explanation cannot reliably govern execution.

Append-Only Record

An Append-Only Record is a chronological record in which new entries may be added, but prior entries are not overwritten, deleted, silently modified, or replaced.

The purpose of append-only structure is not convenience.

It is evidentiary preservation.

TA-14 requires append-only recordkeeping because admissible execution depends on the ability to review what was known, when it was known, how it was preserved, what changed, and whether any action was justified at the time it occurred.

Append-only records preserve sequence.

They preserve dispute.

They preserve uncertainty.

They preserve correction history.

They preserve the difference between original evidence and later interpretation.

A corrected record must not erase the original condition. Correction must itself become part of the chronology.

Reconstruction

Reconstruction is the creation of a later explanation, timeline, state, or conclusion to fill, replace, repair, or simulate a missing or incomplete record.

TA-14 does not treat reconstruction as equivalent to admissible evidence.

A reconstructed state may help explain what likely happened.

It may assist investigation.

It may support human review.

It may inform future safeguards.

But reconstruction cannot silently become admissible state for execution.

The core rule is simple:

Reconstruction may explain.

Reconstruction may not govern execution as if it were origin-preserved truth.

Where a system relies on reconstruction, that reliance must be disclosed, bounded, and prevented from being treated as equivalent to admissible state unless explicitly permitted by the governing standard for a limited purpose.

Imputation

Imputation is the insertion of estimated, modeled, inferred, or statistically filled values into a record where actual evidence is absent.

TA-14 distinguishes imputed data from captured evidence.

Imputation may be useful for analysis.

It may be useful for visualization.

It may be useful for research.

It may be useful for non-executional modeling.

But imputed data cannot be treated as origin evidence unless clearly labeled and prohibited from supporting execution beyond its allowed scope.

If a system fills a gap and then hides the fact that the gap existed, the record has been structurally compromised.

In TA-14, a gap remains a gap.

A filled gap is not the same as an unbroken record.

Drift

Drift is a change between the state that justified action and the state that exists at commit time.

Drift may occur in evidence, environment, authority, scope, configuration, identity, timing, boundary conditions, user permission, system condition, or operating context.

A system may appear internally consistent while drifting away from the original authority or admissible condition.

TA-14 treats drift as a serious execution risk because action may continue under conditions that are no longer valid.

Drift must be checked before execution.

If drift is detected, execution must be blocked, paused, or escalated unless the drift is explicitly permitted within the action’s governing boundary.

No system may assume that a previously valid state remains valid simply because no one has objected.

Authority

Authority is the valid right of a person, system, institution, role, process, or governed layer to rely on a record, bind an action, enforce a decision, or permit execution.

Authority is not the same as access.

A system may have access to data without authority to act.

A person may have permission to view a record without authority to bind execution.

An AI system may generate a recommendation without authority to execute.

A control system may be technically capable of acting while lacking admissible authority to act.

TA-14 requires authority to be explicit, bounded, current, and connected to admissible state.

Authority must not be inferred merely from capability.

The fact that a system can act does not mean it is permitted to act.

Authority Drift

Authority Drift is the condition in which the authority being used at execution no longer matches the authority originally established, preserved, or permitted.

Authority drift may occur when governance roles change, permissions expand, records are transferred, systems are integrated, policies are updated, vendors change, AI agents are added, or operational control shifts without preserving the original authority chain.

Authority drift is dangerous because a system may continue to enforce decisions while no longer operating under the same authority condition that made reliance valid.

TA-14 separates execution validity from authority continuity.

An action may appear valid within the immediate system while still relying on an authority structure that has drifted.

Where authority drift is present, execution must not be treated as fully admissible until authority continuity is restored or revalidated.

Binding

Binding is the formal connection between an admissible state and a proposed action.

Binding prevents action from floating free of the evidence that allegedly supports it.

A valid binding must identify:

1. The record or record segment being relied upon;

2. The admissible state established from that record;

3. The authority under which reliance is permitted;

4. The action being proposed;

5. The scope of the action;

6. The time window in which the binding remains valid;

7. The conditions that invalidate the binding;

8. The required outcome record.

Without binding, execution is unbound.

Unbound execution is the central defect TA-14 exists to correct.

Binding Object

A Binding Object is the formal structure that records and preserves the connection between record, admissible state, authority, action, scope, time, enforcement, and outcome.

The Binding Object is not a narrative explanation.

It is not a note.

It is not a general approval.

It is not a human memory of why an action seemed reasonable.

It is a governed object that makes the action traceable to the admissible conditions that permitted it.

A valid Binding Object must be specific enough to prevent substitution.

A binding for one action cannot be reused for another action.

A binding for one time window cannot be silently extended.

A binding for one scope cannot be expanded without revalidation.

A binding based on one authority cannot be transferred to another authority without preserving the authority chain.

Fail-Closed

Fail-Closed means that when required admissibility conditions are absent, expired, uncertain, contradicted, degraded beyond permitted limits, or unverifiable, the system does not proceed as if execution were allowed.

In TA-14, uncertainty does not automatically create permission.

A system must not act simply because it lacks proof that action is unsafe.

The burden is reversed.

The system must have admissible authority to act.

If that authority is not present, execution must be blocked, paused, or escalated.

Fail-closed behavior is essential because systems that fail open convert uncertainty into action.

TA-14 rejects that condition.

ALLOW

ALLOW is the execution outcome produced when the required admissibility, continuity, authority, binding, scope, and commit-time enforcement conditions are satisfied.

ALLOW does not mean the action is wise in every possible sense.

ALLOW means the action is permitted under the TA-14 execution governance standard because the required state-to-action conditions are valid at commit time.

ALLOW must be recorded.

The basis for ALLOW must be traceable.

The conditions under which ALLOW occurred must be reviewable.

BLOCK

BLOCK is the execution outcome produced when one or more required conditions for admissible execution are absent, invalid, expired, contradicted, out of scope, reconstructed, unbound, or authority-defective.

BLOCK is not failure.

BLOCK is governance functioning correctly.

A system that blocks invalid execution is preserving integrity.

TA-14 treats blocked execution as an important outcome because it proves that action is not automatically privileged over truth.

Where execution cannot be justified by admissible state, blocking is the correct result.

ESCALATE

ESCALATE is the execution outcome produced when a system cannot determine whether execution should be allowed or blocked within the governing authority of the automated or procedural boundary.

Escalation may occur when evidence is degraded, authority is unclear, conditions conflict, human judgment is required, or execution would exceed the permitted scope of the current system.

ESCALATE does not mean proceed.

ESCALATE means the system lacks sufficient authority to resolve the condition on its own.

Escalation must preserve the record of uncertainty.

It must not erase the reason automatic execution was not permitted.

Outcome Record

The Outcome Record is the preserved record of what happened after an execution decision.

It must record whether execution was allowed, blocked, escalated, completed, interrupted, failed, reversed, or otherwise modified.

The Outcome Record is not a substitute for commit-time admissibility.

Post-execution documentation cannot cure invalid execution.

A system cannot act first and make the action admissible later.

Outcome recording serves review, accountability, verification, learning, audit, and future continuity.

It does not replace the requirement that execution be admissible before action occurs.

Reliance

Reliance is the act of using a record, state, conclusion, decision, or authority condition as the basis for action.

TA-14 governs reliance because reliance is where evidence becomes operational.

Many systems collect information.

Fewer systems preserve it.

Fewer still define when it may be relied upon.

TA-14 requires the right to rely to be earned through preserved evidence, continuity, admissibility, authority, and binding.

A record may be visible without being reliable.

A reading may be accurate without being admissible.

A conclusion may be persuasive without being action-permitting.

Reliance must be governed.

Right-to-Rely Boundary

The Right-to-Rely Boundary defines the limit within which a record, state, or authority condition may be used to justify action.

This boundary may be temporal, spatial, procedural, technical, legal, institutional, environmental, or operational.

A record may support reliance within one boundary and fail outside another.

For example, a record may support awareness but not intervention.

A record may support maintenance review but not automated control.

A record may support a single action but not a sequence of additional actions.

The right to rely is not unlimited.

It must be bounded and preserved.

Admissibility

Admissibility is the condition under which evidence, record state, continuity, authority, and scope are sufficient to permit reliance for a defined purpose.

Admissibility does not mean perfection.

It means the record has satisfied the governing criteria required for the proposed use.

TA-14 treats admissibility as a structural condition, not a rhetorical claim.

A system cannot declare something admissible merely because it is useful.

A record cannot be made admissible by confidence alone.

A conclusion cannot be made admissible by expert preference.

Admissibility requires preserved conditions.

Where those conditions are missing, the system must limit reliance.

Admissibility Criteria

Admissibility Criteria are the rules used to determine whether a record or state may be relied upon for a specific action.

These criteria may include:

1. Origin capture;

2. Timestamp integrity;

3. Append-only preservation;

4. Continuity;

5. Identity or source validity;

6. Scope alignment;

7. Authority continuity;

8. Absence of unauthorized reconstruction;

9. Valid binding;

10. Commit-time confirmation;

11. Outcome recordability.

The criteria must be known before execution.

A system cannot change admissibility criteria after action in order to justify what already occurred.

Record Layer

The Record Layer is the layer responsible for preserving what happened.

It observes, captures, timestamps, and preserves evidence.

It does not diagnose.

It does not optimize.

It does not control.

It does not recommend action.

It does not alter the system being evidenced.

The Record Layer must remain separated from interpretation and execution because evidence loses independence when the same layer that records reality also acts upon it.

In TA-14, the Record Layer is foundational because no admissible execution can exist without preserved truth.

Interpretation Layer

The Interpretation Layer is the layer that analyzes, compares, explains, classifies, or derives meaning from the record.

Interpretation may be useful.

Interpretation may be necessary.

Interpretation may guide human understanding.

But interpretation is not the record itself.

TA-14 requires interpretation to remain downstream of the record.

Interpretation must not overwrite the record.

Interpretation must not replace origin evidence.

Interpretation must not be treated as admissible state unless it is explicitly derived, bounded, versioned, traceable, and permitted for the proposed use.

Execution Layer

The Execution Layer is the layer where action occurs.

This may include human action, machine action, AI-agent action, automated control, operational workflow, maintenance intervention, environmental adjustment, institutional decision, or system command.

The Execution Layer must not be permitted to act merely because it can act.

It must be governed by admissible state.

The Execution Layer is where the risk of unbound action becomes real.

For this reason, TA-14 requires execution to pass through a non-bypassable boundary.

Non-Bypassable Enforcement

Non-Bypassable Enforcement means that the execution boundary cannot be skipped, ignored, overridden silently, routed around, or treated as optional.

A governance system that can be bypassed is not an execution governance system.

It is advisory.

TA-14 requires enforcement because the purpose is not merely to describe valid execution.

The purpose is to prevent invalid execution.

If a system allows action to proceed outside the admissibility boundary, the boundary is not structurally valid.

Unbound Action

Unbound Action is action that proceeds without a valid, preserved, and enforceable connection to admissible state.

Unbound action may occur even when people are well-intentioned.

It may occur even when data exists.

It may occur even when a report is accurate.

It may occur even when the action produces a good outcome.

TA-14 does not define validity by outcome alone.

An action is not made admissible because it worked.

An action is admissible only if the conditions permitting it were valid before execution.

Unbound action is the central structural defect TA-14 was created to address.

Environmental Integrity Governance

Environmental Integrity Governance is the structural governance layer that preserves environmental truth as an independent record before interpretation, optimization, control, reporting, or action.

It establishes that environmental conditions must be preserved as evidence, not merely consumed as temporary data.

Environmental Integrity Governance separates observation from action.

It preserves the atmospheric record.

It defines admissibility.

It protects the right-to-rely boundary.

It creates the conditions under which environmental truth can govern execution.

Atmospheric Integrity Record

An Atmospheric Integrity Record is the continuous, append-only, time-sequenced record of atmospheric conditions within a bounded place, system, structure, or environment.

It preserves what was true in the atmosphere over time.

It may include environmental variables, system state, intervention history, continuity conditions, source references, confidence boundaries, and admissibility status.

The Atmospheric Integrity Record is not merely a sensor feed.

It is the governed chronology that allows atmospheric truth to become reviewable and potentially admissible.

Personal Atmospheric Integrity Record

A Personal Atmospheric Integrity Record is the governed, append-only, time-sequenced record of an individual’s environmental exposure across places, transitions, durations, and conditions.

It connects personal exposure to environmental context while preserving separation between observation and interpretation.

The Personal Atmospheric Integrity Record does not diagnose.

It does not assign causality at capture.

It preserves the record of exposure so that later interpretation, review, guidance, or governance may occur from a preserved chronology rather than reconstructed memory.

TA-14

TA-14 is the evidence-locked governance architecture that connects origin-captured truth to admissible action.

It is not a monitoring platform.

It is not a diagnostic method.

It is not a sensor product.

It is not an optimization tool.

It is not a general training slogan.

TA-14 defines the structural conditions under which reality is observed, preserved, admitted, bound, enforced, executed, and recorded.

TA-14 exists because systems often act without admissible truth.

TA-14 makes admissible truth a precondition of execution.

Closing Rule

The terms on this page are controlled because the architecture depends on them.

A system that logs data is not automatically admissible.

A system that recommends action is not automatically governed.

A system that uses AI is not automatically intelligent in the TA-14 sense.

A system that acts quickly is not automatically valid.

A system that produces good outcomes is not automatically admissible.

Under TA-14, the governing question is always:

Was the action bound to admissible truth at the moment execution occurred?

If the answer cannot be proven, execution cannot be treated as admissible.