Intention Registry Entry v1.0

Registry ID: INT-004

Title: Verify

Family: Epistemic

Status: Draft

Version: 1.0.0

Primary function: Test whether a claim, model, state, interpretation, relationship, or process remains valid under evidence, recurrence, comparison, and stress.

0. Canonical Definition

Verify is the intention to test whether a claim, model, state, interpretation, relationship, or process corresponds sufficiently with observable evidence and remains coherent under appropriate conditions.

Verify moves beyond visibility, inquiry, and clarification into active testing.

Reveal makes relevant state visible.

Question opens uncertainty.

Clarify defines what is being examined.

Verify tests whether the resulting claim or model holds.

Verification may concern:

• factual claims
• causal models
• reported events
• system state
• identity claims
• boundaries
• agreements
• operational behavior
• source reliability
• technical function
• restoration outcomes
• organizational mission
• compatibility
• safety
• recurrence
• symbolic or interpretive consistency

Verify does not require absolute certainty.

Its purpose is to establish whether confidence is justified at the level necessary for the current decision, interaction, or system function.

1. Foundational Principle

A claim becomes reliable only when it remains coherent under appropriate evidence, comparison, recurrence, and challenge.

A visible claim may still be false.

A clear explanation may still be incomplete.

A plausible model may fail under different conditions.

An intention may be sincere while its operating assumptions remain invalid.

Verify therefore asks:

Does the claim correspond with the available state?

Does the model predict or explain recurrence?

Does the result survive comparison and stress?

Does confidence match evidence?

Its basic movement is:

claim
→ test
→ comparison
→ confidence adjustment
→ provisional validation or rejection

Verification protects systems from both credulity and permanent suspicion.

Its aim is not to doubt everything.

Its aim is to determine what deserves reliance.

2. Directional Function

Starting condition

Verify commonly arises when:

• a claim has been made
• a model is guiding action
• evidence is incomplete or contested
• a source’s reliability is uncertain
• a result must be reproduced
• operational behavior may diverge from declaration
• safety depends on accuracy
• a boundary or agreement is disputed
• a restoration process claims completion
• a system reports success without sufficient audit
• an interpretation has become influential
• a causal relationship is assumed
• contradictory evidence exists
• recurrence may reveal hidden instability
• confidence exceeds demonstrated validity
• a high-impact decision requires stronger evidence

Intended direction

Verify seeks movement from:

claim
→ tested claim

assumption
→ evidence-weighted conclusion

reported state
→ observed state

single occurrence
→ recurrence-tested pattern

plausibility
→ justified confidence

declared operation
→ audited operation

uncertain source
→ provenance-weighted source

Verify does not seek:

perfect certainty

It seeks:

sufficiently justified confidence for the relevant function

Intended outcome region

A coherent Verify intention seeks an outcome region in which:

• the claim has been tested against relevant evidence
• source provenance is understood
• confidence is proportional to support
• contradictory evidence is included
• assumptions are visible
• relevant uncertainty is localized
• results can be reproduced or independently examined where possible
• failure conditions are known
• limits of validity are explicit
• the system can decide whether to rely, revise, reject, or continue testing
• temporal recurrence supports or weakens the claim

Protected invariants

Verify should preserve:

• truthfulness
• evidence integrity
• source integrity
• uncertainty
• contextual relevance
• repeatability where applicable
• independence of review
• proportionality
• due process
• identity boundaries
• the distinction between absence of proof and proof of absence
• the distinction between provisional validation and permanent truth
• the distinction between testing a claim and attacking the claimant

3. Active Intention Profile

aim: "Determine whether a claim, model, state, interpretation, relationship, or process remains sufficiently valid under relevant evidence, comparison, recurrence, and stress."

target: "A claim, source, model, event, state, process, agreement, relationship, outcome, interpretation, or reported condition."

scope: "Limited to the degree of testing necessary for the decision, risk, responsibility, or system function involved."

horizon: "Active until confidence becomes sufficiently justified, the claim is rejected, the result remains indeterminate, or further testing no longer provides proportionate value."

priority: "Often follows Reveal, Question, or Clarify and may precede Learn, Protect, Coordinate, Restore, Reconcile, Transform, or institutional action."

forbidden_outcomes:
  - "permanent suspicion"
  - "manufactured doubt"
  - "evidence suppression"
  - "burden shifting"
  - "testing without closure"
  - "identity reduction"
  - "predetermined invalidation"
  - "false certainty"

completion_condition: "The claim or model has been tested sufficiently for confidence to be responsibly increased, reduced, withheld, or transferred to further inquiry."

confidence_requirements: "Confidence must scale with evidence quality, provenance, recurrence, independence, and relevance."

reversibility_requirements: "Verification conclusions should remain revisable when new evidence, changed conditions, or failed recurrence materially alter the evidentiary field."

4. Admissibility

Preconditions

Verify is admissible when:

• a meaningful claim or model exists
• the claim can be tested directly or indirectly
• verification serves a legitimate purpose
• the risk or consequence justifies the testing effort
• evidence standards can be defined
• the actor is willing to revise confidence
• contradictory evidence will be considered
• the process does not presume the result
• the testing scope is proportionate
• the subject is not being reduced to a single test result

Supporting evidence

Verification may be supported by:

• direct observation
• primary-source records
• repeatable measurement
• independent confirmation
• internal consistency
• cross-source agreement
• longitudinal recurrence
• stress testing
• audit trails
• technical logs
• chain-of-custody records
• experimental replication
• comparative outcomes
• falsifiable predictions
• documented commitments
• observed operational behavior
• source provenance

Consent conditions

Consent is generally required when verification involves:

• personal data
• intimate identity
• medical testing
• private communications
• bodily access
• confidential relationships
• invasive monitoring
• protected records
• high-risk experimentation
• identity authentication beyond legitimate need

Verification of public claims, institutional actions, fiduciary responsibilities, safety-critical systems, or material impacts may not depend on the voluntary consent of the system being audited.

Even then, the process should preserve:

• proportionality
• legal authority
• scope limitation
• source protection
• procedural fairness
• appeal and correction
• relevant privacy

Boundary conditions

Verify must distinguish:

testing a claim

from:

placing the entire system under permanent suspicion

and:

requiring evidence

from:

demanding impossible proof

Verification should preserve:

• relevance
• defined evidence standards
• temporal limits
• audience boundaries
• procedural symmetry
• response rights
• privacy
• source integrity
• distinction between claimant and claim
• the possibility of closure

Capacity conditions

A system enacting Verify should possess sufficient capacity to:

• define the claim precisely
• identify suitable evidence
• preserve records
• test alternative explanations
• recognize confirmation bias
• compare source quality
• communicate uncertainty
• reproduce or independently review results
• distinguish failed verification from proof of falsehood
• revise confidence
• stop when sufficient testing is complete
• repair harm caused by incorrect conclusions

Scale conditions

Verification should scale according to:

• consequence severity
• number of affected systems
• irreversibility
• public reach
• power asymmetry
• recurrence
• uncertainty
• cost of error
• availability of independent review

A low-risk personal claim may require simple confirmation.

A safety-critical technical system may require repeated, independent, adversarial testing.

A public institution may require transparent and recurring audit.

Null conditions

Verify returns ∅ when:

• no meaningful claim has been defined
• the proposed test cannot affect confidence
• the process presumes failure or guilt
• evidence standards are impossible or intentionally shifting
• testing is being used to delay necessary action
• the cost of verification exceeds the legitimate value
• a less invasive method is sufficient
• the actor refuses contradictory evidence
• the subject has already been sufficiently verified for the current purpose
• verification is actually punishment, harassment, surveillance, or control
• the test measures a proxy unrelated to the claim
• the system lacks capacity to interpret the result
• immediate protective action must precede further testing

5. Intention Provenance

Common sources

Verify may emerge from:

• uncertainty
• due diligence
• scientific inquiry
• safety review
• audit
• contradiction
• source comparison
• public accountability
• technical testing
• restoration review
• legal process
• relational trust repair
• model validation
• risk management
• memory confirmation
• identity authentication
• quality assurance

Legitimate provenance

Verify may be coherently:

• self-authored by a system testing its own assumptions
• invited by a system seeking confirmation
• co-authored through mutual review
• role-derived through audit, science, safety, or oversight
• evidence-derived when claims conflict
• risk-derived where error carries significant cost
• restoration-derived where repair claims require validation
• trust-derived where reliability must be re-established
• implementation-derived where a system must demonstrate function

High-risk provenance

Verify becomes high-risk when driven by:

• permanent suspicion
• coercive surveillance
• ideological purity testing
• status competition
• institutional self-protection
• desire to discredit
• impossible evidence demands
• selective skepticism
• fear of uncertainty
• punitive compliance
• commercial extraction
• public humiliation
• predetermined conclusions

Capture vectors

Suspicion capture

Verification becomes an endless demand for further proof.

Institutional capture

The institution defines the evidence standards and exempts itself from them.

Proxy capture

A convenient metric replaces the actual claim.

Adversarial capture

Testing is designed to cause failure rather than discover validity.

Narrative capture

Only evidence supporting the preferred interpretation is accepted.

Compliance capture

Passing the procedure replaces demonstrating actual coherence.

Authority capture

Institutional endorsement is treated as proof without examining method.

Engagement capture

Controversy is prolonged because uncertainty attracts attention.

6. Interaction Profile

Common interactions

Verify commonly travels through:

• testing
• comparison
• replication
• audit
• source checking
• evidence review
• measurement
• stress testing
• contradiction analysis
• chain-of-custody review
• operational observation
• recurrence tracking
• independent review
• falsification attempt
• model evaluation
• result reproduction

Supporting interactions

Verify is often supported by:

• Reveal
• Question
• Clarify
• Remember
• Learn
• Witness
• Compare
• Audit
• Coordinate
• Protect
• Restore

High-risk interactions

Verify becomes high-risk when paired with:

• force
• surveillance without scope
• public accusation
• impossible proof demands
• repeated testing without closure
• evidence withholding
• selective disclosure
• invasive monitoring
• punitive exposure
• identity authentication beyond need
• adversarial pressure designed to induce failure

Incompatible interactions

Verify is incompatible with:

• evidence fabrication
• predetermined conclusions
• suppression of contradictory data
• shifting standards
• false equivalence
• proxy substitution
• coercive confession
• refusal of independent review
• infinite regress
• identity reduction
• manufactured doubt

7. Operator Profile

Primary operators

Au — Auditability

Makes evidence, process, provenance, and decisions traceable.

Ψ — Presence

Maintains contact with actual state rather than desired conclusions.

Μ — Sensemaking

Interprets evidence through provisional models.

Δ — Differentiation

Separates source from claim, evidence from inference, correlation from cause, and validity from confidence.

Θ — Humility

Keeps confidence proportional and preserves revisability.

Supporting operators

Γ — Selection

Determines which claim, evidence, or test matters.

Ξ — Contrast

Compares prediction with outcome, declaration with operation, or source with source.

Σ — Invariant preservation

Protects truthfulness, fairness, relevance, privacy, and source integrity.

Π — Boundary

Limits scope, access, duration, and audience.

U7 / temporal recurrence

Tests whether apparent validity persists through time.

ℛ — Restoration

Repairs harm caused by false claims, invalid conclusions, or failed verification.

Conditional operators

Amplification

May widen testing or review when consequences scale.

Containment

May be required for dangerous, private, or security-sensitive evidence.

Force

May appear only under valid legal or high-risk conditions and must not convert verification into confession or punishment.

Excluded uses

Verify must not be implemented through operators used to:

• manufacture failure
• suppress contradictory evidence
• impose impossible proof
• convert uncertainty into guilt
• maintain permanent surveillance
• shift the burden unfairly
• treat a test result as total identity
• conceal invalid methods
• redefine the standard after the result

8. Expected State Effects

State-effect caution

Verification may reduce confidence without reducing knowledge quality.

high confidence + weak evidence
→ testing
→ lower confidence + stronger epistemic integrity

Likewise, failed replication does not always prove the original claim false.

It may reveal:

• contextual dependence
• incomplete method
• hidden variables
• measurement error
• insufficient sample
• unstable conditions

9. Intention Across Time

Immediate — I₀

Immediate Verify tests a specific claim or observation.

Examples:

• confirming a record
• checking a source
• testing whether a boundary was crossed
• reproducing a result
• comparing an action with an agreement
• checking whether a system is functioning

Tactical — I₁

Tactical Verify examines reliability across multiple interactions.

Examples:

• monitoring recurrence
• conducting an audit
• comparing independent sources
• reviewing system logs
• testing a model under several conditions
• validating a restoration process
• evaluating whether behavior matches stated intention

Strategic — I₂

Strategic Verify creates durable validation architecture.

Examples:

• scientific replication systems
• institutional audits
• safety certification
• public accountability
• continuous technical monitoring
• recurring legitimacy review
• adversarial testing
• source-provenance systems
• longitudinal outcome measurement
• versioned validation protocols

Identity-trajectory conflict

Verify conflicts with identity trajectory when testing becomes:

• permanent suspicion
• inability to trust
• compulsive rechecking
• refusal to act
• identity reduction
• constant surveillance
• impossible proof demands
• denial of valid testimony
• inability to recognize sufficient evidence

The trajectory conflict is:

Verify as justified reliance
⊥
Verify as refusal to trust

10. Completion and Exit

Completion condition

Verify is complete when:

• the claim has been tested against relevant evidence
• confidence has been responsibly adjusted
• source provenance is sufficiently understood
• major contradictions have been addressed
• limits of validity are explicit
• the result is sufficient for the current decision
• further testing offers diminishing legitimate value
• the next intention can proceed

Exit condition

Active verification should end when:

• sufficient evidence exists for provisional reliance or rejection
• the testing standard has been met
• further verification becomes repetitive or invasive
• irreducible uncertainty must be carried
• the question belongs to Learn, Protect, Restore, Coordinate, or another intention
• additional testing would delay necessary action
• the process begins reproducing suspicion rather than knowledge

Common transition

Verify commonly transitions into:

Verify → Learn
Verify → Protect
Verify → Refuse
Verify → Coordinate
Verify → Restore
Verify → Reconcile
Verify → Transform
Verify → Preserve
Verify → Release

Recurrence condition

Verify may be reactivated when:

• new evidence appears
• conditions change
• prior results fail to recur
• implementation diverges
• source reliability changes
• hidden variables emerge
• risk increases
• a previously validated system begins producing anomalous outcomes
• the claim expands beyond its original scope

Non-completion warning

Verify without a sufficiency threshold becomes permanent suspicion, surveillance, or refusal to act.

11. Polarity Architecture

Coherent form — Evidence-Based Validation

The coherent form of Verify:

• defines the claim clearly
• uses relevant evidence
• preserves contradictory data
• supports independent review
• adjusts confidence
• recognizes limits
• permits closure
• remains revisable
• protects boundaries
• connects testing to actual decisions

Coherent Verify establishes warranted reliance without pretending to eliminate all uncertainty.

Deficit form — Unverified Reliance

Too little Verify produces:

• credulity
• false confidence
• rumor propagation
• unsafe systems
• unsupported authority
• repeated error
• mission drift
• fraudulent claims
• inability to identify failure
• trust without evidence
• reliance on status or appearance
• restoration declared without outcome testing

Excess form — Compulsive Rechecking

Too much Verify produces:

• paralysis
• duplication
• exhaustion
• delay
• resource waste
• inability to trust
• repeated demands for proof
• permanent monitoring
• erosion of autonomy
• refusal to accept provisional validity
• endless testing after standards are met

Shadow inversion — Permanent Suspicion

Verify inverts when testing no longer seeks valid confidence and instead exists to prevent trust, closure, or action.

Examples:

• shifting proof standards
• repeatedly reopening settled evidence
• treating all claims as deceptive
• requiring impossible certainty
• endless loyalty tests
• permanent surveillance
• rejecting testimony solely because it cannot be externally reproduced
• using uncertainty as proof of guilt
• treating verification failure as total identity failure

Captured form — Selective Verification

Captured Verify applies scrutiny asymmetrically.

Examples:

• demanding proof from low-power systems while accepting institutional claims
• auditing workers while leadership remains opaque
• questioning user testimony while trusting platform metrics
• testing only outcomes that threaten a preferred narrative
• accepting friendly sources without review
• applying strict standards to opponents and weak standards to allies
• validating procedure while ignoring actual harm

False form — Compliance Theater

False Verify creates the appearance of testing without meaningfully examining validity.

Examples:

• checkbox audits
• certifications without real review
• metrics disconnected from outcomes
• staged demonstrations
• internal review with no independence
• tests optimized to pass
• security theater
• procedural compliance treated as system safety
• replication that omits critical conditions
• dashboards that confirm only expected results

False Verify preserves confidence without earning it.

12. Failure Modes

Formation failure

• Verification is selected from distrust rather than genuine uncertainty.
• The actor presumes deception.
• A preferred result determines the test.
• Risk is exaggerated to justify surveillance.
• The test is designed around an irrelevant proxy.
• Institutional status substitutes for evidence.
• The actor cannot tolerate provisional uncertainty.

Scope failure

• Testing expands beyond the claim.
• A single failed result defines total identity.
• Temporary review becomes permanent monitoring.
• Private state is examined without relevance.
• Verification requirements spread to unrelated systems.
• A local result is generalized beyond its valid domain.

Selection failure

• Verify is used when Reveal is needed.
• Testing replaces immediate protection.
• More evidence is demanded where sufficient proof already exists.
• Verification delays restoration.
• A relational issue requiring trust is treated only as an audit problem.
• A non-testable meaning question is forced into measurement.

Operator failure

• Evidence and inference are conflated.
• Auditability is asymmetric.
• Contradictory data is discarded.
• Humility is absent.
• Replication conditions differ materially.
• The metric does not represent the claim.
• The standard changes after results appear.
• Source provenance is ignored.

Boundary failure

• Consent is bypassed.
• Monitoring becomes invasive.
• Refusal is treated as guilt.
• Private data is retained indefinitely.
• Identity is reduced to authentication data.
• Testing continues after completion.
• The subject cannot inspect or challenge the process.

Completion failure

• No amount of evidence is sufficient.
• Tests are endlessly repeated.
• New standards are added after each result.
• Verification cannot transition into decision.
• The system preserves uncertainty because it benefits from continued review.
• Provisional validation is never permitted.

Validation failure

• Passing the test is treated as proof of total safety.
• Certification is treated as permanent validity.
• Procedure is mistaken for outcome.
• One successful result is generalized.
• Failed recurrence is ignored.
• The test is never evaluated against real-world effects.
• Internal review is treated as independent.

Attribution failure

• Failed verification is treated as proof of deception.
• Lack of evidence is treated as evidence of absence.
• Inability to reproduce is treated as proof of fabrication.
• A wrong claim is treated as total identity corruption.
• Requests for verification are assumed to be hostile.
• Institutional endorsement is treated as proof of good intention.

13. Forbidden Outcomes

Verify must not produce:

• permanent suspicion
• endless testing
• invasive surveillance
• manufactured doubt
• impossible proof requirements
• identity reduction
• evidence suppression
• burden shifting
• selective standards
• predetermined invalidation
• procedural theater
• delay of necessary protection or repair
• false certainty
• punishment for indeterminate results

14. Declared, Operational, and Inferred Intention

Declared signals

Declared Verify commonly appears through statements such as:

• “We need to confirm this.”
• “What evidence supports the claim?”
• “Can the result be reproduced?”
• “Does the behavior match the stated policy?”
• “What would show that this model is wrong?”
• “Has this held across time?”
• “Can another system independently review it?”
• “What is the source provenance?”

Declared verification language does not itself prove that the process is fair or open.

Operational signals

Verify is operationally present when:

• the claim is defined clearly
• evidence standards are stated in advance
• contradictory evidence is preserved
• source quality is evaluated
• confidence changes with results
• independent review is possible
• the process permits both validation and rejection
• scope is bounded
• closure conditions exist
• methods are auditable
• the same standards apply across power levels
• real-world outcomes are compared with test results

Outcome signals

Verify has produced a coherent outcome when:

• confidence is more justified
• unreliable claims are weakened or rejected
• valid claims become more dependable
• uncertainty becomes localized
• source reliability is clearer
• system decisions improve
• hidden failure conditions become visible
• trust is better calibrated
• restoration or protection can proceed on stronger evidence
• recurrence supports the conclusion

Common misattributions

• Treating verification as distrust.
• Assuming failed verification proves dishonesty.
• Treating lack of proof as proof of absence.
• Assuming institutional certification guarantees validity.
• Treating one test as universal proof.
• Assuming more testing always increases truth.
• Treating skepticism as inherently rigorous.
• Assuming testimony is invalid because it is not externally reproducible.
• Treating procedural compliance as coherent operation.
• Assuming a verified claim remains permanently valid.

15. Responsibility Profile

Responsibility analysis for Verify should ask:

• Was verification necessary?
• Was the claim defined accurately?
• Were evidence standards relevant and known in advance?
• Could the test genuinely validate or reject the claim?
• Were contradictory results included?
• Was the process independent?
• Were privacy and consent preserved?
• Were power asymmetries considered?
• Was the burden of proof assigned fairly?
• Did the system recognize limits?
• Was sufficient evidence allowed to produce closure?
• Did testing delay protection or repair?
• Were real-world outcomes compared with the test?
• Were errors corrected?
• Did the system revise confidence?
• Did the verification process itself produce harm?
• Was recurrence examined?

16. Restoration Profile

Restoration triggers

Restoration is required when:

• verification becomes permanent suspicion
• evidence is selectively included
• standards shift after results
• surveillance exceeds legitimate scope
• false invalidation produces harm
• a proxy replaces the actual claim
• procedural compliance conceals failure
• independent review is blocked
• testing continues after completion
• indeterminate results are framed as guilt
• private information is unnecessarily exposed
• verification delays necessary repair

Cessation requirements

• stop unnecessary testing
• end disproportionate surveillance
• suspend invalid metrics
• cease repeating settled reviews
• stop treating indeterminate results as failure
• halt publication of unsupported conclusions
• separate legitimate verification from punishment
• move into action when evidence is sufficient

Audit requirements

The audit should determine:

• what claim was tested
• who defined the standard
• whether the method matched the claim
• which evidence was included or excluded
• whether the process was independent
• whether standards shifted
• who bore the testing burden
• whether privacy was violated
• who benefited from continued uncertainty
• whether closure was possible
• whether test results matched real outcomes
• whether correction pathways existed

Repair actions

Repair may require:

• correcting the conclusion
• reopening review under valid standards
• restoring access or rights lost through false invalidation
• ending unnecessary monitoring
• deleting improperly retained data
• acknowledging methodological failure
• compensating for material harm
• correcting public records
• accepting previously dismissed evidence
• redesigning the verification process
• validating actual outcomes rather than proxies

Authorship restoration

Authorship restoration may include:

• allowing the affected system to inspect the evidence
• enabling response and appeal
• recognizing first-person testimony where relevant
• separating failed claims from total identity
• restoring control over private data
• ending compulsory re-verification
• returning authority after the verification threshold is met

Completion recovery

To restore completion:

• restate the precise claim
• define the relevant standard
• identify sufficient evidence
• remove irrelevant tests
• document uncertainty
• make a provisional conclusion
• establish review triggers
• transfer the result to Learn, Protect, Restore, or Coordinate
• stop testing when the legitimate threshold is reached

Recurrence prevention

• publish standards before testing
• require independent review
• audit proxy validity
• define closure conditions
• preserve contradictory evidence
• version methods
• limit data retention
• require proportionality
• separate verification from punishment
• test real-world outcomes
• review asymmetrical evidence demands
• require periodic rather than continuous revalidation where appropriate

17. Intention Relationships

Reinforcing intentions

INT-001 — Reveal

Reveal makes relevant evidence visible for testing.

INT-002 — Question

Question identifies which claim or assumption requires verification.

INT-003 — Clarify

Clarify defines the claim, variables, and evidence standards.

INT-005 — Remember

Remember provides historical state for recurrence and consistency testing.

INT-006 — Learn

Learn integrates verification results into improved capacity.

INT-012 — Preserve

Verify can determine whether a structure remains worthy of preservation.

Tension intentions

INT-013 — Connect

Repeated verification can support trust or prevent connection when no evidence is ever sufficient.

INT-014 — Invite

An invitation may require openness rather than immediate proof.

INT-015 — Witness

Some states must first be received before being externally tested.

INT-018 — Reconcile

Verification may support accountability but can also delay relational repair.

INT-021 — Stabilize

Testing may expose instability or become destabilizing if continual.

INT-023 — Liberate

Verification can expose capture or become another mechanism of control.

Common predecessors

• INT-001 — Reveal
• INT-002 — Question
• INT-003 — Clarify
• INT-005 — Remember
• INT-015 — Witness

Common successors

• INT-006 — Learn
• INT-007 — Protect
• INT-008 — Refuse
• INT-012 — Preserve
• INT-017 — Coordinate
• INT-018 — Reconcile
• INT-019 — Transform
• INT-020 — Restore
• INT-021 — Stabilize
• INT-023 — Liberate

18. Domain Applications

Personal and relational systems

Verify may involve:

• confirming whether an agreement was understood
• comparing words with repeated behavior
• checking whether a boundary remains respected
• confirming shared expectations
• testing whether trust is being restored
• distinguishing isolated error from recurring pattern

Example:

A person states that a harmful pattern has ended.

Verification does not require permanent surveillance.

It may involve observing whether behavior changes consistently across time, whether responsibility is accepted, and whether repair continues without repeated prompting.

Organizational systems

Verify may support:

• internal audit
• policy compliance
• mission–operation comparison
• financial review
• risk validation
• safety testing
• performance evaluation
• quality assurance
• restoration review
• incentive analysis

Organizational Verify becomes false when compliance documentation is accepted while actual outcomes remain unexamined.

Governance

Verify supports governance through:

• independent oversight
• election audits
• public records
• judicial review
• budget review
• policy evaluation
• impact measurement
• conflict-of-interest examination
• procedural transparency
• recurring legitimacy checks

Governance verification must apply to the systems exercising power, not only those subject to it.

Artificial intelligence

Verify in AI systems may include:

• checking retrieved facts
• validating tool output
• comparing generated claims with sources
• testing whether a requested action completed
• confirming user intent before irreversible operations
• evaluating model uncertainty
• checking whether safety conditions remain active
• testing outputs against specified constraints

AI Verify should not repeatedly ask users to reconfirm information already available or shift all validation responsibility onto them.

Healthcare

Verify may include:

• confirming diagnosis
• repeating abnormal tests
• checking treatment response
• validating medication lists
• comparing reported symptoms with longitudinal state
• confirming consent
• monitoring recurrence

Healthcare Verify should not dismiss lived testimony merely because a single test fails to reproduce it.

Justice

Verify may concern:

• evidence authenticity
• chain of custody
• testimony consistency
• forensic methods
• procedural compliance
• institutional conduct
• legal identity
• remedy completion

Justice requires that state evidence and methods remain open to meaningful challenge.

Education

Verify supports:

• knowledge assessment
• source checking
• model testing
• peer review
• experiment replication
• citation review
• self-correction
• understanding checks

Educational Verify should test understanding rather than reward memorized compliance alone.

Symbolic systems

Verify may test:

• whether a claimed historical meaning is documented
• whether a symbolic interpretation fits the context
• whether patterns recur across cultures
• whether an inversion claim is supported
• whether a symbol’s operational use matches its stated meaning
• whether primary and later meanings have been distinguished

Symbolic verification should not erase layered meaning simply because one interpretation is easier to document.

Game systems

Verify may operate through:

• testing NPC claims
• comparing evidence
• recurring-world-state checks
• validating symbolic combinations
• confirming restored relationships
• testing whether a world rule is genuine
• comparing visible and hidden system variables

A coherent game implementation should allow verification to change trust, available actions, and interpretation rather than functioning only as a pass/fail gate.

19. Compact Reference Card

INT-004 — VERIFY

Family:
Epistemic

Aim:
Test whether a claim, model, state, interpretation, relationship, or process remains valid under relevant evidence, comparison, recurrence, and stress.

Target:
A claim, source, model, event, process, agreement, outcome, interpretation, or reported condition.

Preserves:
Evidence integrity, uncertainty, context, privacy, fairness, revisability, and proportionality.

Common interactions:
Testing, comparison, replication, audit, source checking, measurement, recurrence tracking, and independent review.

Primary operators:
Au, Ψ, Μ, Δ, Θ.

Completion:
Confidence is responsibly increased, reduced, withheld, or transferred after sufficient testing.

Forbidden outcomes:
Permanent suspicion, endless testing, surveillance, manufactured doubt, impossible proof, selective standards, and false certainty.

O⁺:
Evidence-based validation.

O⁻:
Permanent suspicion.

Deficit:
Unverified reliance.

Excess:
Compulsive rechecking.

Captured form:
Selective verification.

False form:
Compliance theater.

Canon line:
Verify tests whether confidence is deserved and whether a claim remains coherent under evidence, comparison, recurrence, and stress.

20. Canon Lockbox

• Verify is an intention, not an operator.
• Verify tests a claim; it does not define the entire claimant.
• Evidence standards should be stated before results are known.
• Confidence must scale with evidence.
• Verification must preserve contradictory data.
• Source authority does not replace source examination.
• Passing a procedure does not prove coherent operation.
• Failed verification does not automatically prove deception.
• Absence of evidence is not always evidence of absence.
• One successful test does not guarantee universal validity.
• One failed test does not always establish total invalidity.
• Verification must remain relevant to the claim.
• Proxies must not replace actual state.
• Verification must permit closure.
• Verification without closure becomes permanent suspicion.
• Verification without symmetry becomes domination.
• Verification without outcome testing becomes compliance theater.
• Verification without humility becomes false certainty.
• Verification without boundaries becomes surveillance.
• Temporal recurrence determines whether validity persists beyond the initial test.

21. Canon Line

Verify tests whether confidence is deserved by examining whether a claim, model, or state remains coherent under relevant evidence, comparison, recurrence, and stress without converting uncertainty into permanent suspicion.