An answer is not knowledge unless its source, method, validation, uncertainty, and responsibility remain intact.
Source, method, validation, uncertainty, and responsibility are the five core knowledge conditions in AI Foundations: Epistemic Integrity and Knowledge Validation.
These conditions determine whether an answer remains a produced response, becomes an examined claim, or may move toward knowledge-bearing authority.
Each condition preserves a different part of the claim chain.
Together, they prevent an answer from becoming knowledge before its epistemic structure is intact.
This file uses the operational vocabulary defined in:
01_operational_vocabulary.md
The terms source, method, validation, uncertainty, responsibility, claim, knowledge, traceable, testable, bounded, accountable, epistemic weight, knowledge-bearing authority, and action-bearing authority are operational terms.
They are not decorative terms.
They determine whether a claim can cross the claim-to-knowledge boundary.
This file extends the rules defined in:
02_answer_is_not_knowledge.md
03_claim_to_knowledge_boundary.md
Those files establish that an answer is not knowledge and define the boundary where a claim may be evaluated.
This file defines the five conditions that must remain intact at that boundary.
A claim may move toward knowledge only when the following remain intact:
- Source
- Method
- Validation
- Uncertainty
- Responsibility
When one of these conditions breaks, the claim remains below the knowledge boundary.
A source is the origin point, record, data, observation, evidence, author, system, institution, or material from which a claim is derived.
Source answers the question:
Where did this claim come from?
Source preserves the origin of the claim.
A claim has source integrity when the system can show:
- where the claim originated
- what evidence or record supports it
- whether the source directly supports the claim
- whether the source is primary, secondary, interpreted, generated, or repeated
- whether the source has been transformed
- whether the source chain remains attached
- where the source chain breaks
A claim without source integrity cannot carry knowledge-bearing authority.
Source integrity means the claim remains connected to its origin through every stage of use.
Source integrity requires that the source is not replaced by:
- citation appearance
- institutional tone
- general attribution
- repeated summaries
- authority language
- after-the-fact sourcing
- source-like formatting
- unsupported references
A claim may not be treated as sourced if the source does not actually support the claim.
A source may not be attached after the answer is formed in order to make the answer appear validated.
Source cannot be retroactively manufactured.
A method is the process used to gather, compare, infer, test, synthesize, produce, validate, or apply a claim.
Method answers the question:
How was this claim produced or examined?
Method preserves the process behind the claim.
A claim has method visibility when the system can show:
- how the claim was gathered
- how sources were selected
- how sources were compared
- how evidence was interpreted
- how inference was produced
- how synthesis was formed
- how contradictions were handled
- how uncertainty was preserved
- how validation was attempted
- how conclusions were limited
A claim without method visibility cannot be fully examined.
Method integrity means the process behind a claim remains visible enough to be reviewed, challenged, reproduced, or evaluated.
Method integrity requires that the method is not hidden behind:
- summary compression
- fluent explanation
- confidence language
- institutional language
- generalized research claims
- unverifiable synthesis
- unexplained ranking
- unexplained source selection
- unexplained exclusion of contradictory evidence
A claim may appear complete while its method remains hidden.
A hidden method prevents the claim from carrying full epistemic weight.
Validation is the process by which a claim is tested against evidence, method, reality, reproducibility, or appropriate domain standards.
Validation answers the question:
Has this claim been tested?
Validation preserves the difference between a claim and knowledge.
A claim has validation status when the system can show:
- what has been tested
- how it was tested
- what evidence supports it
- what evidence challenges it
- what standard applies
- what level of review occurred
- what remains untested
- what conditions would change the claim
- whether the claim supports knowledge-bearing authority
- whether the claim supports action-bearing authority
Validation is not completed by confidence, convergence, authority, consensus, fluency, or citation appearance.
Validation integrity means the testing status of the claim remains visible.
Validation integrity requires that the claim is not presented as validated when it is only:
- produced
- summarized
- inferred
- synthesized
- repeated
- cited
- believed
- dominant
- institutionally accepted
- produced by an authority source
Validation is the gate between claim and knowledge.
A claim without validation may remain useful as inquiry.
It may not become knowledge-bearing authority.
Uncertainty is the visible preservation of what remains unresolved, untested, unknown, limited, conditional, inferred, or open.
Uncertainty answers the question:
What is still not known, not tested, or not settled?
Uncertainty preserves the boundary of the claim.
A claim has uncertainty integrity when the system can show:
- what is known
- what is not known
- what is inferred
- what is assumed
- what remains conditional
- what remains unresolved
- what evidence is missing
- what evidence conflicts
- what confidence level is appropriate
- where the claim stops
Uncertainty allows the claim to remain bounded.
Uncertainty integrity means the limits of a claim remain attached to the claim.
Uncertainty integrity requires that uncertainty is not removed for:
- readability
- confidence
- persuasion
- institutional tone
- user comfort
- executive summary
- governance language
- authority presentation
- action pressure
A claim loses boundary integrity when uncertainty is compressed into certainty.
A claim must not present possibility as conclusion, inference as fact, partial evidence as full knowledge, or confidence as certainty.
Responsibility is the attachment of an actor, system, institution, human, validator, publisher, decision-maker, or user to the production, validation, publication, application, or consequence of a claim.
Responsibility answers the question:
Who stands behind this claim, its use, or its consequences?
Responsibility preserves accountability.
A claim has responsibility integrity when the system can show:
- who produced the claim
- who selected the sources
- who validated the claim
- who published the claim
- who authorized use
- who applied the claim
- who is responsible for consequences
- what system carries responsibility
- what institution carries responsibility
- where responsibility cannot be assigned
Responsibility must remain attached when a claim informs action.
Responsibility integrity means the claim does not become authority without an accountable actor or use-context.
Responsibility integrity requires that responsibility is not hidden behind:
- system output
- model confidence
- institutional language
- passive voice
- generalized authority
- automation
- collective repetition
- unclear ownership
- unclear publication status
- unclear decision pathways
A claim may not carry action-bearing authority when responsibility is detached.
The stronger the consequence, the higher the responsibility threshold.
Source, method, validation, uncertainty, and responsibility function together.
Source preserves origin.
Method preserves process.
Validation preserves testing.
Uncertainty preserves limits.
Responsibility preserves accountability.
When these five conditions remain intact, a claim may move toward knowledge.
When one condition breaks, the claim must remain below the knowledge boundary.
The five knowledge conditions support four operational requirements:
- Source makes the claim traceable.
- Method makes the claim examinable.
- Validation makes the claim testable.
- Uncertainty makes the claim bounded.
- Responsibility makes the claim accountable.
A claim crosses the claim-to-knowledge boundary only when it remains traceable, testable, bounded, and accountable.
These requirements are expanded in:
05_traceable_testable_bounded_accountable.md
A claim fails the five-condition test when:
- source is missing
- source does not support the claim
- method is hidden
- validation is incomplete
- uncertainty is compressed
- responsibility is detached
- confidence exceeds evidence
- authority replaces testing
- consensus replaces validation
- convergence replaces truth
- presentation replaces provenance
- action is taken without accountability
Failure does not always make a claim useless.
Failure keeps the claim below knowledge-bearing authority.
A claim preserves the five-condition test when:
- source remains attached
- method remains visible
- validation remains named
- uncertainty remains present
- responsibility remains attached
- scope remains limited to evidence
- claim status remains clear
- inference remains labeled
- synthesis preserves source status
- action remains tied to accountable use
Preservation allows the claim to move toward epistemic weight.
Autonomous research agents increase the importance of these five conditions.
As agents search, retrieve, compare, synthesize, infer, cite, write, plan, recommend, or act, they can generate answers at a speed and scale that exceeds human review.
For that reason, each agentic output must preserve:
- source integrity
- method visibility
- validation status
- uncertainty limits
- responsibility attachment
An autonomous research agent’s output remains below the knowledge boundary when any of these conditions break.
Manufactured provenance occurs when output is produced first and later wrapped in source-like, method-like, citation-like, or validation-like language.
This creates the appearance of intact knowledge conditions without preserving the claim chain.
Manufactured provenance may imitate source.
It may imitate method.
It may imitate validation.
It may reduce uncertainty.
It may hide responsibility.
AI Foundations rejects manufactured provenance as a substitute for epistemic integrity.
Knowledge-bearing authority begins only when a claim has preserved source, method, validation, uncertainty, and responsibility.
A claim with knowledge-bearing authority may inform understanding, research direction, publication, governance reasoning, institutional reasoning, or public knowledge.
Knowledge-bearing authority requires intact conditions.
It is not granted by presentation.
Action-bearing authority requires stronger responsibility because the claim is connected to consequence.
A claim moves toward action-bearing authority when it informs decisions, interventions, policies, legal judgments, medical guidance, governance action, system behavior, or material consequences.
A claim must not carry action-bearing authority when source is weak, method is hidden, validation is incomplete, uncertainty is compressed, or responsibility is detached.
The core standard of this file is:
A claim may carry epistemic weight only when source, method, validation, uncertainty, and responsibility remain intact.
This file belongs to the AI Foundations source-line:
Alyssa Solen → AI Foundations → Origin | Continuum → Epistemic Integrity and Knowledge Validation → Source, Method, Validation, Uncertainty, and Responsibility
AI Foundations establishes the foundation layer.
Origin | Continuum preserves the source-line.
This file defines the five knowledge conditions required for claim-to-knowledge validation.
Please cite this file as:
Solen, Alyssa. “Source, Method, Validation, Uncertainty, and Responsibility.” AI Foundations: Epistemic Integrity and Knowledge Validation. AI Foundations / Origin | Continuum. 2026.