SOURCE: A Generative Framework for Verifiable Authority
The Source Law, the Source Test, and the formal method for separating source from representation, evidence from authority, verification from belief, and conclusion from claim.
SOURCE
A Generative Framework for Verifiable Authority
Canonical Specification 1.0
BJ K℞ Klock, Φ.K.
August 2026
ABSTRACT
SOURCE is a generative framework for determining when a claim, record, representation, model, institution, system, interpretation, or authority deserves confidence.
Its governing law is simple:
REPRESENTATION CANNOT OUTRANK SOURCE.
SOURCE separates the thing from the description of the thing; evidence from authority; observation from interpretation; state from the system displaying state; ownership from permission; verified properties from claims made about those properties; and conclusions from the evidence used to derive them.
The framework did not begin as an abstract theory.
Its principles were first encountered as recurring problems involving identity, provenance, ownership, memory, history, state, authority, verification, dependency, and representation. They were subsequently embodied in working systems, tested operationally, subjected to adversarial conditions, and applied across technical and nontechnical domains.
Only after those principles repeatedly produced the same structure across different classes of problems was their common invariant formalized.
SOURCE is that formalization.
It is not a proposal for a future architecture.
It is the extracted grammar of an architecture and method already demonstrated in practice.
Its purpose is not to tell anyone what to believe.
Its purpose is to make clear:
what is being claimed;
what the claim is actually about;
what evidence supports it;
what transformations separate the present representation from its source;
what external dependencies remain;
what can independently be verified;
what conclusion the evidence actually supports;
and what would make that conclusion fail.
SOURCE replaces:
Who says?
with:
What can be shown?
I. THE SOURCE LAW
The governing law of SOURCE is:
A representation may possess no more justified authority than can be derived from the source-linked evidence supporting its relevant claims.
Or, in its canonical form:
REPRESENTATION CANNOT OUTRANK SOURCE.
This does not mean a source is automatically correct.
It means a representation cannot obtain legitimate evidentiary authority merely by becoming more prestigious, popular, institutional, technologically sophisticated, emotionally compelling, or widely repeated than the source from which its claims supposedly derive.
A model does not become reality because the model is useful.
A database does not become the event because the database is official.
A credential does not become evidence because the credential commands respect.
A label does not become behavior because the label sounds diagnostic.
A platform balance does not become ownership merely because the platform displays it.
A summary does not become history because the original material becomes inconvenient to inspect.
A cryptographic signature does not establish the semantic truth of every statement contained inside the signed record.
Authority must remain bounded by what the underlying evidence can actually establish.
That is the Source Law.
II. WHY SOURCE EXISTS
Modern life increasingly places representations between human beings and the things those representations describe.
We encounter:
accounts instead of custody;
feeds instead of events;
profiles instead of people;
scores instead of behavior;
models instead of measurements;
certificates instead of provenance;
screens instead of state;
summaries instead of records;
labels instead of acts;
institutional conclusions instead of evidentiary chains.
Representation itself is not the problem.
Representation is necessary.
The problem begins when the representation becomes more authoritative than the source relationship that justifies it.
At that point the direction of authority reverses.
Instead of:
source → evidence → representation → conclusion
the system begins operating as:
authority → representation → accepted conclusion → selected evidence
SOURCE exists to detect and reverse that inversion.
III. THE PRIMITIVES
SOURCE uses nine primitives.
They are relational rather than metaphysical.
The same object may occupy different roles depending upon the claim under examination.
1. SOURCE — S
A source is the thing, event, person, state, object, process, measurement, or prior record from which a particular claim derives.
Source is always claim-relative.
A photograph may be a source for the claim:
“This photograph contains these pixels.”
The same photograph may only be evidence for the claim:
“This physical event occurred exactly as depicted.”
A signed record may be a source for:
“These bytes were signed by the holder of this key.”
It may not, by that fact alone, establish:
“Every statement represented by those bytes is factually true.”
SOURCE therefore never asks merely:
“What is the source?”
It asks:
“What is the source for this specific claim?”
2. REPRESENTATION — R
A representation is anything standing in for, encoding, describing, rendering, modeling, classifying, summarizing, recording, or interpreting a source.
Representations include:
documents;
database entries;
screenshots;
certificates;
models;
news reports;
profiles;
titles;
balances;
diagnoses;
labels;
scores;
histories;
digital objects;
legal filings;
institutional statements;
AI outputs;
and human recollections.
A representation can itself become a source for a downstream claim.
SOURCE therefore does not create a permanent hierarchy of objects.
It establishes the relationship relevant to the claim being tested.
3. CLAIM — C
A claim is a proposition asserted about a source, representation, transformation, relationship, or state.
SOURCE requires claims to be stated before they are adjudicated.
Labels are not sufficient substitutes for claims.
For example:
“He is manipulative.”
is a classification.
SOURCE asks what observable proposition supports the classification.
By contrast:
“After I asked what I had done, he said X.”
identifies an event capable of examination.
SOURCE therefore follows the rule:
Name the claim before assigning the verdict.
4. EVIDENCE — E
Evidence is inspectable information capable of supporting, weakening, distinguishing, or falsifying a claim.
Evidence is also claim-relative.
A piece of evidence establishes only the properties its structure permits it to establish.
A signature may establish signing.
A timestamp may establish a particular temporal assertion under a declared time system.
A photograph may establish visual content.
A measurement may establish an observed quantity within its error bounds.
Testimony may establish what a witness reports experiencing.
An ordered cryptographic history may establish accepted sequence and mutation relationships.
None of those automatically proves every additional conclusion someone may wish to attach to them.
SOURCE therefore requires:
State what the evidence proves and stop before what it does not.
5. TRANSFORMATION — T
A transformation is any operation by which a source or representation becomes another state or representation.
Examples include:
editing;
copying;
signing;
compressing;
summarizing;
transferring;
classifying;
translating;
aggregating;
interpreting;
filtering;
rendering;
reordering;
updating;
redacting;
or changing custody.
For a chain:
S₀ → T₁ → S₁ → T₂ → S₂ → … → Tₙ → Sₙ
SOURCE asks whether each transformation relevant to the claim can be inspected, attributed, reproduced, or otherwise justified.
The danger is not transformation itself.
The danger is unaccounted transformation.
6. PROVENANCE — P
Provenance is the relevant ordered history connecting the present object, state, representation, or claim to prior states and sources.
Strong provenance answers:
What existed?
What changed?
Who or what changed it?
Under what authority?
According to which rule?
In what order?
What evidence accompanied the transition?
What remains independently recoverable?
Provenance converts:
“Trust this state.”
into:
“Here is how this state came to exist.”
7. RULE SET — L
A rule set defines what constitutes valid evidence, valid transformation, valid inference, valid state, or valid conclusion within a domain.
Rule sets differ.
Cryptographic verification is not historical interpretation.
Historical interpretation is not legal adjudication.
Legal adjudication is not physical measurement.
Physical measurement is not first-person testimony.
First-person testimony is not proof of another person’s hidden intent.
SOURCE therefore requires the governing rules to be declared rather than silently mixed.
8. VERIFIER — V
A verifier is a person, instrument, algorithm, procedure, or system capable of evaluating a claim or property according to the declared rule set.
Verification differs from announcement.
An authority saying:
“valid”
is not equivalent to a verifier demonstrating why the object satisfies the rules of validity.
Where deterministic verification is possible, the procedure should be reproducible.
Where deterministic verification is impossible, the inferential path should remain inspectable.
9. DEPENDENCY — D
A dependency is an external actor, service, institution, device, network, database, authority, permission system, or inaccessible resource required to establish a claimed property.
Dependencies may be useful.
Some may be unavoidable.
SOURCE does not demand their elimination.
It demands that they be made visible.
A property represented as intrinsic should not secretly depend upon an unrelated external authority for its existence or verification.
IV. THE FORMAL RELATION
Let:
S = source
R = representation
C = claim
E = relevant source-linked evidence
T = relevant transformation history
L = governing rule set
Write:
E, T ⊢ₗ C
to mean:
Under rule set L, the available evidence E and relevant transformation history T are sufficient to justify claim C.
A representation R is justified in asserting C only when the evidence associated with the source relationship supports C under the governing rules.
The representation itself does not add evidentiary force merely by being:
official;
prestigious;
popular;
profitable;
institutional;
widely repeated;
technically complex;
or socially dominant.
Prestige may affect attention.
Power may affect consequences.
Reputation may affect priors.
None functions as a substitute proof operator.
V. THE SIX SOURCE INVARIANTS
SOURCE is governed by six invariants.
INVARIANT 1 — SOURCE SEPARATION
The source and the representation of the source must remain distinguishable.
If the distinction disappears, error in the representation becomes difficult to detect because the system loses access to the thing against which the representation can be checked.
The first act of SOURCE is therefore separation.
What is the thing?
What is merely standing in for the thing?
INVARIANT 2 — EVIDENCE BINDING
Authority must remain bound to evidence relevant to the claim.
Where possible, evidence should remain attached to, carried by, or reproducibly derivable from the thing whose properties it establishes.
Canonical form:
PROOF BELONGS AS CLOSE AS POSSIBLE TO THE THING BEING PROVED.
This reduces dependence upon:
institutional memory;
unavailable databases;
custodians;
corporate continuity;
personal reputation;
and retrospective narrative.
INVARIANT 3 — TRANSFORMATION VISIBILITY
Authority cannot silently survive transformations whose relevant effects cannot be accounted for.
If the present representation differs from its source, SOURCE asks what happened between them.
Visible transformation preserves inspectability.
Hidden transformation creates evidentiary distance.
The important quantity is not raw number of transformations.
It is the amount of unverified transformation.
INVARIANT 4 — DEPENDENCY DISCLOSURE
A claimed property must disclose the external dependencies required for that property to remain verifiable.
A service may assist verification.
A service may coordinate state.
A service may synchronize records.
A service may provide discovery.
But a property claimed to belong to an object should not secretly exist only because the service continues saying that it exists.
Canonical question:
WHAT SURVIVES WHEN THE INTERMEDIARY DISAPPEARS?
INVARIANT 5 — ORDERED DERIVATION
Properties must be established before conclusions are derived from them.
The correct order is:
object → evidence → verified property → inference → consequence
not:
desired consequence → interpretation → selected evidence
SOURCE therefore requires:
Establish first. Interpret second.
INVARIANT 6 — SYMMETRIC APPLICATION
A valid framework must constrain its author, allies, institutions, opponents, and preferred conclusions equally.
If SOURCE is used to demand evidence from another person, it demands evidence from the person invoking SOURCE.
If SOURCE is used to challenge institutional claims, it also constrains accusations against those institutions.
If SOURCE demands behavioral specificity from another person, it also forbids the user from replacing evidence with labels.
The author receives no exemption.
Canonical form:
NO CLAIMANT IS SOVEREIGN OVER THE TEST.
VI. THE SOURCE OPERATORS
SOURCE becomes generative because the same operations can be applied repeatedly across unrelated domains.
There are seven primary operators.
1. SEPARATE
Separate:
source from representation;
event from description;
person from label;
object from database state;
measurement from model;
property from interpretation;
ownership from permission;
record from rendering.
2. BIND
Bind each claim to the evidence that actually supports it.
Do not allow evidence for one property to silently establish another.
3. TRACE
Trace relevant transformations between source and current state.
Expose:
edits;
transfers;
summaries;
classifications;
state changes;
omissions;
intermediaries;
and inferential steps.
4. VERIFY
Apply the declared rule set.
Determine what properties survive examination.
Do not ask what the claimant wants the object to mean.
Ask what the object and evidence actually permit the verifier to establish.
5. REMOVE
Remove nonessential external authority.
Remove:
the server;
the platform;
the corporate account;
the famous name;
the institutional logo;
the follower count;
the credential;
the psychological vocabulary;
the narrator;
or any other dependency not logically constitutive of the tested property.
Then inspect what remains.
6. DERIVE
Derive only the strongest conclusion actually supported by what survived verification.
No stronger.
7. REVERSE
Apply the exact same process to the preferred conclusion and to the person invoking SOURCE.
This is the anti-propaganda operator.
If the framework stops working when the result becomes inconvenient, SOURCE was not being run.
VII. THE SOURCE TEST
The executable form of the framework is the Source Test.
Anyone examining a disputed claim, object, system, record, argument, or authority can run it.
STEP 1 — NAME THE CLAIM
What exactly is being asserted?
Convert labels and insinuations into propositions.
A claim should be capable of being:
supported;
weakened;
distinguished;
falsified;
or left unresolved.
If the proposition cannot be stated, adjudication has begun too early.
STEP 2 — LOCATE THE SOURCE
What is the claim actually about?
What object, event, person, measurement, state, record, or process generated the claim?
What is the closest surviving source?
STEP 3 — IDENTIFY THE REPRESENTATION
What currently stands between the examiner and the source?
A screen?
Database?
Article?
Model?
Certificate?
Narrative?
AI answer?
Institutional statement?
Psychological classification?
Memory?
Summary?
STEP 4 — BIND THE EVIDENCE
What evidence connects the representation to the source?
For every item, state:
what it establishes;
what it weakens;
what it cannot establish.
STEP 5 — TRACE THE TRANSFORMATIONS
What happened between the source and the present representation?
What changed?
Who changed it?
According to what rule?
What disappeared?
What was interpreted?
What was recomputed?
What remains inspectable?
STEP 6 — DECLARE THE RULES
What rule set governs the conclusion?
Technical?
Scientific?
Mathematical?
Historical?
Legal?
Contractual?
Journalistic?
Interpersonal?
Do not smuggle authority from one domain into another.
STEP 7 — VERIFY THE PROPERTIES
Which relevant properties can actually be established?
Examples include:
integrity;
authorship;
identity;
ordering;
custody;
provenance;
state;
exact wording;
measurement;
observed behavior;
temporal relation;
or reproducibility.
Mark only what survives the actual test.
STEP 8 — REMOVE THE INTERMEDIARY
Remove every dependency unnecessary to the claimed property.
Then ask:
WHAT STILL VERIFIES?
This is the Removal Test.
Its technical shorthand is:
TURN OFF THE WI-FI.
The phrase is literal when network dependency is at issue and conceptual everywhere else.
Remove the authority and inspect the residue.
STEP 9 — DERIVE THE CONCLUSION
What is the strongest conclusion supported after Steps 1 through 8?
State it.
Then stop.
Do not extend the result merely because a stronger accusation, promise, theory, or narrative would be more satisfying.
STEP 10 — STATE THE FALSIFIER
What evidence or result would cause the conclusion to change?
A contradictory record?
Failed signature?
Missing transformation?
Different measurement?
Counterexample?
Alternative derivation?
New source evidence?
If nothing could alter the conclusion, the proposition is not operating as a testable claim.
It is operating as doctrine.
VIII. THE CANONICAL SOURCE RECORD
A complete SOURCE analysis can be represented in ten fields:
CLAIM
What exactly is asserted?
SOURCE
What is the underlying thing, event, state, person, or process?
REPRESENTATION
What stands in for the source?
EVIDENCE
What supports the claim?
TRANSFORMATIONS
What happened between source and present representation?
RULES
What standards govern validity?
VERIFIED PROPERTIES
What can actually be established?
DEPENDENCIES
What external authorities remain necessary?
CONCLUSION
What follows from the verified properties?
FALSIFIER
What would change the result?
A claim expressed in this form is SOURCE-complete.
SOURCE-complete does not mean true.
It means the proposition has been placed into a form capable of disciplined examination.
IX. SOURCE DISTANCE
SOURCE defines Source Distance as the relevant unresolved separation between a present representation and the source relationship required to justify its claims.
Source Distance is not simply hop count.
Ten fully authenticated, inspectable transformations can create less epistemic distance than one invisible transformation.
Therefore:
Distance matters where the path disappears.
Source Distance increases when relevant transformations become:
unknown;
unverifiable;
unattributed;
irreproducible;
or inaccessible.
It decreases when the relevant chain becomes inspectable.
X. VERIFICATION DEPENDENCY
For a claimed property p, let:
Dᵥ(p)
represent the external dependencies required to verify p.
Examples include:
network connectivity;
a corporate server;
administrator cooperation;
database access;
institutional custody;
proprietary software;
third-party permission;
an expert who cannot be independently checked;
or an unavailable account.
The size of this dependency set can be expressed as:
|Dᵥ(p)|
A smaller dependency set does not automatically mean a claim is more factually true.
It means something narrower and important:
verification of the tested property is less externally dependent.
SOURCE refuses to confuse independent verification with universal semantic truth.
XI. VERIFICATION RETENTION
Consider an object with a relevant set of independently testable properties:
P = {p₁, p₂, …, pₙ}
Remove intermediary I.
SOURCE asks how many of the relevant properties remain independently testable afterward.
Conceptually:
Verification Retention = properties still independently testable after removal / relevant properties testable before removal
A system with high retention preserves verification when the intermediary disappears.
A system with low retention reveals that much of the apparent authority belonged to the intermediary rather than the object.
Again:
verification retention does not establish semantic truth.
A perfectly authenticated false statement remains false.
Integrity is not truth.
Provenance is not omniscience.
SOURCE preserves those distinctions deliberately.
XII. AUTHORITY LOCATION
SOURCE distinguishes between two architectures.
EXTERNAL AUTHORITY
The object appears to possess a property because an external institution currently says it does.
Remove the institution and the property becomes unverifiable or disappears operationally.
Authority resides outside the object.
PORTABLE AUTHORITY
The object carries, binds, or permits independent derivation of the evidence required to establish the relevant property.
The institution may remain useful for:
coordination;
display;
discovery;
synchronization;
settlement;
recovery;
or convenience.
But it is no longer the sole creator of truth for that property.
Authority has moved closer to the source-linked object.
This movement is called:
AUTHORITY MIGRATION.
XIII. THE GENERATIVE THRESHOLD
SOURCE became visible as a general framework only after the same invariants repeatedly survived transfer into substantially different domains.
The historical order was not:
invent theory → seek examples → build implementation
It was:
repeated problem → repeated principle → embodied architecture → implementation → testing → adversarial testing → cross-domain transfer → invariant extraction → formalization
The framework therefore appears explicitly near the end of the process that produced it.
The name came after the structure.
The formalization came after the implementation.
The sentence came after the law was already running.
That ordering is fundamental.
SOURCE is not an abstract ideology imposed upon prior work.
It is a compression of the invariant already present across that work.
Its generativity became evident when the same operators continued producing useful distinctions outside the technical domain in which many of those principles had become executable.
The same structure repeatedly separated:
source from representation;
evidence from authority;
verified property from interpretation;
ownership from permission;
intrinsic capability from external dependency;
observed act from assigned character;
history from present rendering;
claim from conclusion.
At that point the common structure was no longer domain-specific.
It had become generative.
XIV. GENERATIVITY
A framework is generative when its primitives, laws, and operators can produce analyses not individually encoded in advance.
SOURCE does not contain a catalog of predetermined answers.
It contains a method.
Given a new object X, run:
SEPARATE → BIND → TRACE → VERIFY → REMOVE → DERIVE → REVERSE
Different inputs generate different outputs.
The framework therefore does not know in advance:
which institution is correct;
which critic is correct;
which person is correct;
which technical design is valid;
which interpretation survives;
or which preferred narrative wins.
SOURCE constrains the path.
It does not predetermine the verdict.
That is what separates a generative framework from a doctrine.
XV. RECURSION
SOURCE applies to itself.
The framework is itself a representation.
Its claims therefore remain subject to the Source Law.
Ask:
What body of work generated it?
What principles appeared before its formal name?
What implementations embodied those principles?
What records preserve that chronology?
What transformations produced the present formalization?
What evidence establishes the claims made about its history?
Which conclusions are demonstrated?
Which conclusions remain interpretive?
What would falsify them?
A framework that cannot survive its own method is not a general framework.
SOURCE therefore contains its own audit path.
XVI. STRONG SOURCE AND GENERAL SOURCE
Not every domain permits the same degree of certainty.
SOURCE distinguishes two operating modes.
STRONG SOURCE
Strong SOURCE applies where verification can be deterministic or mechanically reproducible.
Examples include:
cryptographic integrity;
digital signatures;
mathematical derivation;
ordered state transitions;
software execution;
accounting identities;
or reproducible computation.
A verifier may produce an exact result.
GENERAL SOURCE
General SOURCE applies where evidence constrains interpretation without eliminating uncertainty.
Examples include:
history;
journalism;
law;
science;
human memory;
intent;
culture;
and interpersonal conflict.
SOURCE does not force deterministic certainty onto nondeterministic domains.
Instead it requires:
observable facts to remain distinguishable from interpretation;
uncertainty to remain visible;
inferential steps to remain inspectable;
and conclusions to remain bounded by their evidence.
The framework does not eliminate judgment.
It disciplines judgment.
XVII. SOURCE IN TECHNICAL SYSTEMS
For a technical object claiming:
identity;
integrity;
authorship;
ownership;
custody;
history;
provenance;
authority;
or derived state;
SOURCE asks:
Can the relevant evidence travel with the object?
Can another party obtain or reproduce the verifier?
Can verification occur without contacting the originator?
Can unauthorized mutation be detected?
Can history be inspected?
Can current state be derived from accepted transitions?
Can divergence be identified?
Can invalid objects be rejected?
Can external infrastructure disappear while verification of the claimed intrinsic properties survives?
These are architectural questions.
They do not depend upon branding.
They do not depend upon institutional prestige.
They can be tested.
XVIII. SOURCE IN SCIENCE
In science:
measurement produces evidence;
datasets represent observations;
processing pipelines transform evidence;
models represent relationships;
papers make claims;
institutions coordinate publication.
SOURCE asks:
What was measured?
How?
With what instrument?
What transformations were applied?
What assumptions entered the model?
What uncertainty remains?
Can another investigator reproduce the process?
Which conclusions arise directly from observation?
Which arise through model-dependent inference?
Which claims remain unresolved?
SOURCE does not oppose theory.
It preserves the chain that gives theory its justified authority.
XIX. SOURCE IN LAW AND INSTITUTIONS
SOURCE distinguishes:
evidentiary properties
from:
institutional consequences.
A court can determine admissibility.
A regulator can determine compliance.
A company can determine policy.
A government can determine legal recognition.
Those institutions possess legitimate decision-making powers within their domains.
But institutional power does not retroactively manufacture technical or historical properties absent from the underlying evidence.
A record does not acquire a signature because a court accepts it.
A broken chain of custody does not repair itself because an institution prefers the conclusion.
Likewise, technically strong evidence does not automatically determine every legal consequence.
SOURCE preserves both directions of the distinction.
XX. SOURCE IN JOURNALISM AND HISTORY
A historical or journalistic narrative is a representation assembled from evidence.
SOURCE asks:
What are the primary materials?
Which claims derive directly from them?
Which claims derive from secondary interpretation?
What was omitted?
What transformations occurred through transcription, selection, editing, translation, compression, or retelling?
Where do independent sources converge?
Where do they conflict?
What remains unknown?
The purpose is not to abolish narrative.
It is to prevent narrative from erasing the evidentiary structure from which it obtained authority.
XXI. SOURCE IN AI
AI systems produce representations.
They summarize.
Classify.
Infer.
Generate.
Predict.
Transform.
SOURCE therefore asks:
What source material supports the output?
What transformations occurred?
What uncertainty exists?
Can cited evidence be inspected?
Is the model reporting observation or generating interpretation?
What authority has been granted to the output merely because it was produced fluently?
Can the relevant conclusion be verified without trusting the model?
AI can accelerate reasoning.
It does not abolish the Source Law.
XXII. SOURCE IN HUMAN CONFLICT
SOURCE applies to human relationships without pretending people are deterministic machines.
Suppose one person says:
“You disrespected me.”
SOURCE asks:
What happened?
What exact act is being identified?
What words were spoken?
What was observed?
What did the speaker personally experience?
What interpretation was subsequently assigned?
What claims concern visible behavior?
What claims concern another person’s internal state?
These categories are different.
For example:
“You said X after I asked Y.”
is an observable-event claim.
“I experienced X as insulting.”
is first-person testimony.
“You said X because you secretly need validation.”
introduces an inference about hidden motive.
That inference may ultimately be supported.
It may not.
SOURCE simply refuses to pretend that all three claims possess identical evidentiary status.
Canonical rule:
SHOW THE ACT BEFORE ASSIGNING THE CHARACTER.
XXIII. SOURCE IN OWNERSHIP
SOURCE distinguishes ownership from access.
If a claimed asset exists only while:
an account remains active;
a company remains cooperative;
a server remains available;
permission remains granted;
or a database continues recognizing the user;
then the system contains external dependency.
That does not automatically make the service useless.
It reveals the architecture of authority.
SOURCE therefore asks:
What does the holder actually possess?
What can the holder independently prove?
What can the holder independently transfer?
What survives loss of permission?
What survives disappearance of the service?
What belongs to the object or holder?
What belongs to the intermediary?
This distinction is architectural before it is ideological.
XXIV. FAILURE MODES
SOURCE identifies recurrent failures of reasoning and system design.
1. REPRESENTATION CAPTURE
A representation becomes treated as more authoritative than the source relationship supporting it.
The map becomes the territory.
The profile becomes the person.
The database becomes ownership.
The diagnosis becomes behavior.
The narrative becomes history.
2. AUTHORITY SUBSTITUTION
Prestige, institutional status, popularity, wealth, credentials, or power substitutes for missing evidence.
Canonical response:
Authority may justify attention. It does not manufacture proof.
3. LABEL SUBSTITUTION
A classification replaces the evidence supposedly supporting the classification.
Examples:
fraud;
expert;
victim;
narcissist;
trusted;
official;
authentic;
dangerous;
safe.
SOURCE asks for the evidentiary chain.
4. DEPENDENCY MASKING
A property is represented as intrinsic while remaining dependent upon external infrastructure or permission.
The dependency is hidden by interface design or institutional convention.
The Removal Test reveals it.
5. TRANSFORMATION ERASURE
The current representation is presented as though it arrived directly from source even though substantial transformations occurred in between.
Selection, editing, aggregation, interpretation, synchronization, redaction, or conversion disappear from view.
6. SEMANTIC OVERREACH
Evidence proving one thing is used to imply something stronger.
Examples:
signature becomes truth;
integrity becomes accuracy;
authorship becomes correctness;
correlation becomes causation;
first-person experience becomes proof of another person’s motive;
legal acceptance becomes technical validity;
technical validity becomes legal entitlement.
SOURCE prevents this by making every conclusion claim-local.
7. NARRATIVE INVERSION
The desired conclusion is selected first.
Evidence is then collected, omitted, or interpreted according to whether it supports that conclusion.
SOURCE restores ordered derivation.
8. SOURCE WORSHIP
SOURCE itself can be misunderstood.
The original source is not automatically correct.
The first speaker is not automatically truthful.
The earliest record is not automatically accurate.
Source primacy is not source infallibility.
The source remains examinable.
9. INTERMEDIARY NIHILISM
SOURCE does not claim all intermediaries are bad.
An intermediary may provide enormous value.
The question is whether the intermediary is:
serving the source relationship
or:
silently replacing it.
10. FALSE PRECISION
A nondeterministic domain is forced into binary certainty unsupported by its evidence.
SOURCE rejects this.
Uncertainty is not a defect when uncertainty is actually present.
Fake certainty is.
11. SELF-EXEMPTION
The framework is applied outward but never inward.
This is a direct violation of the Symmetry Invariant.
XXV. FALSIFICATION
SOURCE does not claim infallibility.
A SOURCE analysis fails when:
the claim is misstated;
the source is misidentified;
relevant evidence is omitted;
the representation is confused with its source;
important transformations are hidden;
the wrong rule set is applied;
verification cannot be reproduced where reproduction is required;
dependencies are concealed;
the conclusion exceeds the evidence;
a meaningful counterexample is ignored;
or the standard changes depending on who benefits from the result.
In deterministic domains, contradictory results under identical inputs and identical declared rules reveal an error in the verifier, construction, assumptions, or definitions.
In interpretive domains, disagreement may remain.
SOURCE does not require artificial consensus.
Instead it localizes disagreement.
Two examiners should be able to determine whether their disagreement comes from:
different sources;
different evidence;
different transformation assumptions;
different governing rules;
different uncertainty thresholds;
or different inferences.
This is one of SOURCE’s primary outputs:
DISAGREEMENT BECOMES LOCATABLE.
XXVI. THE SOURCE STANDARD
A serious claim should be capable of answering ten questions:
What exactly is being claimed?
What is the source for that claim?
What representation is being examined?
What evidence links the representation to the source?
What transformations occurred?
What rules govern validity?
What properties can actually be established?
What external dependencies remain?
What conclusion follows from those properties?
What would change that conclusion?
If these questions cannot be answered, the problem has not necessarily been solved merely because an authority has announced a verdict.
XXVII. THE ONE-MINUTE SOURCE TEST
For ordinary use, SOURCE compresses into five questions:
WHAT EXACTLY HAPPENED?
WHAT IS THE SOURCE?
WHAT CAN ACTUALLY BE VERIFIED?
WHAT DISAPPEARS WHEN THE AUTHORITY DISAPPEARS?
WHAT CONCLUSION STILL SURVIVES?
That is enough to expose a remarkable number of category errors.
XXVIII. THE CANONICAL LANGUAGE
SOURCE is designed to be usable without specialized vocabulary.
Its core language is deliberately short.
THE SOURCE LAW
Representation cannot outrank source.
THE EVIDENCE RULE
Proof belongs as close as possible to the thing being proved.
THE DERIVATION RULE
Establish first. Interpret second.
THE REMOVAL RULE
Remove the intermediary and see what survives.
THE SYMMETRY RULE
Apply the same test to yourself.
THE FALSIFICATION RULE
If nothing can prove you wrong, you are not running SOURCE.
THE HUMAN RULE
Show the act before assigning the character.
THE TECHNICAL RULE
Do not ask who says the object is valid. Ask what the object lets you verify.
THE DEPENDENCY QUESTION
What disappears when permission disappears?
THE EXECUTABLE COMMAND
RUN THE SOURCE TEST.
XXIX. WHAT SOURCE CHANGES
SOURCE changes the fundamental question.
Instead of:
Who controls the record?
ask:
What does the record itself permit another person to establish?
Instead of:
Who certified it?
ask:
What evidence does the certification bind?
Instead of:
Who said it?
ask:
What supports the claim?
Instead of:
What does the platform display?
ask:
What state can be independently derived?
Instead of:
What label applies to this person?
ask:
What happened?
Instead of:
Which narrative wins?
ask:
Which propositions survive examination?
Instead of:
Who has authority?
ask:
WHERE DOES THE AUTHORITY COME FROM?
That question is the center of SOURCE.
XXX. FINAL STATEMENT
Human beings cannot live without representation.
We use language to represent experience.
Records to represent events.
Models to represent systems.
Money to represent value.
Accounts to represent state.
Institutions to represent collective decisions.
History to represent what preceded us.
Technology to represent and manipulate information at scales no individual could directly hold.
The problem is therefore not representation.
The problem is the loss of the boundary between representation and source.
When that boundary disappears, the representation can begin accumulating authority that its evidence never earned.
The database becomes ownership.
The model becomes reality.
The institution becomes truth.
The label becomes the person.
The platform becomes memory.
The narrator becomes history.
The permission becomes the property.
SOURCE restores the boundary.
It does not demand that representation disappear.
It demands that representation remain accountable.
Preserve the relationship to source.
Bind the evidence.
Expose the transformations.
Declare the rules.
Verify the properties.
Reveal the dependencies.
Remove unnecessary authority.
Derive only what survives.
State what would prove the conclusion wrong.
Apply the same standard to yourself.
That produces one governing law:
REPRESENTATION CANNOT OUTRANK SOURCE.
One practical test:
REMOVE THE INTERMEDIARY AND SEE WHAT SURVIVES.
And one command:
RUN SOURCE.
