Source Provenance Triangulation¶
Essence¶
Source Provenance Triangulation is the discipline of asking what each source is, where it came from, how close it is to the matter being reconstructed, what perspective it carries, whether it is independent of other sources, and how much confidence its evidence can support. It is useful whenever an account could become persuasive simply because it is coherent, frequently repeated, or backed by citations that have not been examined.
The archetype is not just “collect more sources.” It is a way to keep claims attached to their evidentiary grounds. A claim supported by a firsthand log, a later memoir, an official statement, an interview, and a scholarly synthesis should not be treated as five identical pieces of support. Each source occupies a different evidentiary position and must be interpreted accordingly.
Compression statement¶
When reconstructing events or conditions from mixed sources, Source Provenance Triangulation prevents evidentiary flattening by classifying source type, recording provenance, assessing proximity and bias, checking independent corroboration, and attaching confidence labels to claims.
Canonical formula: claim confidence = f(source type, provenance, proximity, perspective, independence, corroboration, unresolved gaps)
When to Use This Archetype¶
Use this archetype when a decision, story, investigation, retrospective, or explanation depends on evidence from multiple kinds of sources. It is especially important when sources conflict, when an account will become institutional memory, when a finding affects accountability, or when a polished secondary account may be standing on weak primary evidence.
It is also appropriate when source type alone is too crude. A primary source can be biased, partial, edited, translated, or far from the relevant event. A secondary source can be derivative, circular, or carefully triangulated. The archetype helps reviewers avoid both naive trust and indiscriminate skepticism.
Structural Problem¶
The structural problem is evidentiary flattening. Sources are placed behind claims as if they all carry the same kind of support. A citation, log, interview, official record, memory, artifact, dataset, or later analysis may all appear in the same reference pool, but each has different access, incentives, proximity, and transformation history.
When those differences are hidden, weak claims can harden into consensus. A claim repeated by many sources may still have one origin. A vivid witness account may dominate a dry but more direct record. An official report may have strong provenance but narrow perspective. A later synthesis may be clear while distancing the reader from the original evidence.
Intervention Logic¶
The intervention begins by naming the claim or account that needs support. Reviewers then classify the sources, record provenance, assess proximity, profile perspective and bias, map interpretive layers, compare independent corroboration, and attach confidence labels. The result is not merely a pile of evidence; it is a claim-support structure that preserves why each claim is strong, weak, contested, or provisional.
The core logic is comparative. A source is not evaluated in isolation. Its evidentiary meaning changes when it is compared against other sources with different origins, positions, and vulnerabilities to error. Agreement matters most when sources are independent; disagreement matters when it reveals perspective, timing, omitted context, or genuine uncertainty.
Key Components¶
Source Provenance Triangulation prevents evidentiary flattening by keeping each source's evidentiary position attached to the claims it supports, rather than treating diaries, logs, official records, interviews, and later syntheses as interchangeable proof. The pattern starts with Source Type Classification, which distinguishes firsthand, contemporaneous, derivative, interpretive, summary, testimonial, material, and analytical sources — not to rank them automatically, but to identify what evidentiary work each can perform. The Provenance Record then captures origin, authorship, custody path, transformations, access conditions, and unresolved gaps, so reviewers know who produced what, when, for whom, and what edits or losses occurred along the way. Event Proximity assesses closeness to the event under review across time, space, role, access, and instrumentation, distinguishing a temporally close but narrow source from a distant but contextually rich one. A Perspective Bias Profile documents the source's position, incentives, audience, and blind spots — not to disqualify it, but to locate it.
The remaining components turn these per-source assessments into a comparative judgment about claim strength. The Interpretive Layer Map separates observation from transcription, translation, selection, commentary, and synthesis, exposing where meaning changed and whether a claim rests on original evidence or inherited framing. The Corroboration Check then compares sources for independent agreement, disagreement, silence, and shared dependence — because multiple sources copied from one origin should not be mistaken for independent confirmation, and agreement matters most when sources are differently vulnerable to error. The Confidence Label carries the resulting reasoned judgment with the claim, explaining whether confidence is high, moderate, low, contested, or provisional, and citing source type, provenance, proximity, perspective, independence, corroboration, and missing evidence rather than presenting a naked score. Together the components convert evidence use into an inspectable claim-support map that can be revised when new sources arrive, instead of a polished narrative that hides its own weaknesses.
| Component | Description |
|---|---|
| Source Type Classification ↗ | Source type classification distinguishes firsthand, contemporaneous, derivative, interpretive, summary, testimonial, material, and analytical sources. The goal is not to rank sources automatically, but to identify what kind of evidentiary work each source can perform. A sensor log, a diary, an interview, a newspaper article, and a later synthesis should not be asked to support the same claim in the same way. |
| Provenance Record ↗ | The provenance record captures origin, authorship or production context, custody path, transformations, access conditions, and unresolved gaps. It answers questions such as: Who produced this? When? For what audience? How was it preserved or transmitted? What edits, translations, selections, or losses occurred before review? |
| Event Proximity ↗ | Event proximity assesses closeness to the event or condition under review. Proximity is not only about time. It includes spatial location, role, access, mediation, technical instrumentation, and whether the source observed the relevant part of the event. A source can be temporally close but conceptually narrow, or temporally distant but valuable for explaining motives and context. |
| Perspective Bias Profile ↗ | The perspective bias profile documents position, incentives, constraints, audience, conflicts of interest, and blind spots. Bias does not mean a source is useless. It means the source is located. The question is what the source could see, what it had reason to emphasize or hide, and what other perspectives are needed to interpret it responsibly. |
| Interpretive Layer Map ↗ | The interpretive layer map separates observation from transcription, translation, selection, commentary, synthesis, and later reuse. This is crucial because derivative sources often present interpretation as if it were evidence. Mapping layers helps reviewers identify where meaning changed and whether a claim rests on original evidence or inherited framing. |
| Corroboration Check ↗ | The corroboration check compares sources for independent agreement, disagreement, silence, contradiction, and shared dependence. Corroboration is strongest when sources are independent and differently vulnerable to error. Multiple sources copied from the same origin may create a false sense of confidence. |
| Confidence Label ↗ | The confidence label carries the resulting judgment with the claim. It should explain whether confidence is high, moderate, low, contested, or provisional, and why. A useful label mentions source type, provenance, proximity, perspective, independence, corroboration, and missing evidence rather than presenting confidence as a naked score. |
Common Mechanisms¶
| Mechanism | Description |
|---|---|
| Source Criticism Protocol ↗ | A source criticism protocol implements the archetype through structured questions about authorship, purpose, audience, context, proximity, and transmission. It is a mechanism, not the archetype itself: the protocol is useful only if it changes how claims are weighted and qualified. |
| Evidence Provenance Log ↗ | An evidence provenance log records origins, custody, transformations, access dates, and gaps. It operationalizes the provenance record component and makes future revision possible. The log is not enough on its own; it must feed into corroboration and confidence judgments. |
| Triangulation Matrix ↗ | A triangulation matrix places claims against sources so agreement, disagreement, independence, and source-type diversity can be seen at once. This mechanism helps prevent a coherent story from hiding unsupported or singly sourced claims. |
| Audit Trail Review ↗ | Audit trail review inspects logs, version histories, document histories, or system traces for gaps and transformations. It is especially useful in incident review, data governance, and institutional accountability contexts where records may have changed after the fact. |
| Chain of Custody Record ↗ | A chain of custody record documents who handled evidence, when, under what conditions, and with what transformations. It implements a custody-focused variant of the archetype. It should not be confused with proof of truth; it mainly supports proof of handling and integrity path. |
| Witness / Source Comparison ↗ | Witness or source comparison evaluates firsthand accounts by role, timing, memory risk, access, incentive, and corroboration against non-testimonial sources. It helps avoid over-weighting the most vivid or authoritative account. |
| Citation Lineage Review ↗ | Citation lineage review traces claims through references and derivative accounts. It detects circular citation, inherited assumptions, and apparent consensus that actually comes from a single weak origin. |
| Confidence Annotation Rubric ↗ | A confidence annotation rubric attaches reasoned confidence levels to claims. It implements the archetype by making claim strength explicit and revisable rather than hidden in prose. |
| Conflicting Source Table ↗ | A conflicting source table keeps contradictions visible. Instead of forcing premature resolution, it records each source’s claim, position, possible explanation, and unresolved question. |
Parameter / Tuning Dimensions¶
The first tuning dimension is evidentiary stakes. A public accountability finding, legal investigation, safety review, or canonical historical account needs stronger provenance and corroboration than a low-stakes internal note.
The second dimension is source diversity. When sources share one institutional origin, one role, one data pipeline, or one citation lineage, reviewers need stronger perspective and independence checks. When sources are diverse but uneven in quality, reviewers need explicit weighting rules.
The third dimension is proximity threshold. Some domains require contemporaneous records; others can use retrospective testimony if memory risk and interpretive layer are acknowledged. The threshold should match the type of claim being made.
The fourth dimension is confidence vocabulary. A lightweight setting may use high, medium, low, and unknown. A formal investigation may need claim-level labels such as well-corroborated, single-source, contested, derivative, inferred, or unsupported.
The fifth dimension is disclosure granularity. Provenance should be transparent enough for review, but source identities, whistleblower details, victim information, or sensitive records may need redaction or access control.
Invariants to Preserve¶
The first invariant is claim-source linkage: material claims should remain tied to the sources that support or challenge them. When this link is broken, accounts become difficult to revise.
The second invariant is source-type distinction. Firsthand, derivative, interpretive, and summary materials must not collapse into one citation pool.
The third invariant is provenance visibility. Origin, transmission, transformation, and access gaps should remain visible because they shape what a source can responsibly support.
The fourth invariant is corroboration independence. Agreement should only increase confidence when sources are not merely copying the same origin or sharing the same hidden dependency.
The fifth invariant is confidence humility. A claim should not sound more settled than its source base warrants.
Target Outcomes¶
The intended outcome is a more reliable account. The archetype does this by making evidentiary support inspectable rather than rhetorical.
A second outcome is visible uncertainty. Readers can tell which claims are strongly supported, which are provisional, which are contested, and which depend on a narrow source base.
A third outcome is reduced evidentiary laundering. Repetition, citation density, official tone, and polished synthesis no longer automatically create confidence.
A fourth outcome is better revision. Because claims carry source links and confidence reasons, future evidence can update the account without starting over.
A fifth outcome is fairer interpretation. Perspective gaps and positional bias become explicit, reducing overreliance on the easiest, loudest, most formal, or most powerful source set.
Tradeoffs¶
The archetype trades speed for rigor. It can slow down reporting, research, postmortems, and institutional learning, but it reduces the risk of preserving errors as settled accounts.
It also trades narrative smoothness for evidentiary clarity. A fully qualified account may be less elegant than a simple story, but it is more honest about what is known and unknown.
It trades inclusion for weighting discipline. More sources can improve coverage, but the archetype still requires unequal weighting when sources differ in proximity, independence, provenance, and corroboration.
Finally, it trades immediate closure for future revisability. The recordkeeping burden is justified when the account will be reused, contested, audited, or used to assign responsibility.
Failure Modes¶
A common failure mode is citation laundering, where derivative sources cite each other until a weak claim looks well established. Citation lineage review and independence checks mitigate this.
Another failure mode is primary-source fetishism. Reviewers may assume firsthand sources are automatically superior, even when they are biased, partial, edited, or narrow. Proximity and perspective analysis prevent this oversimplification.
A third failure mode is false equivalence. The archetype asks for multiple perspectives, but it does not require treating all claims as equally supported. Confidence should remain unequal where the evidence is unequal.
A fourth failure mode is provenance theater. Teams may fill out metadata fields without letting provenance change claims, confidence, or unresolved-gap notes. The mitigation is to require every relevant provenance finding to affect interpretation.
A fifth failure mode is single-score flattening. A reliability score can be useful, but it should not replace reasons. The reasons are the transferable knowledge.
A sixth failure mode is over-triangulation. Not every claim needs exhaustive review. The level of triangulation should scale with stakes, uncertainty, and reuse risk.
Neighbor Distinctions¶
Source Provenance Triangulation is distinct from Traceability Linking. Traceability Linking connects items, records, decisions, and consequences. This archetype uses traceability but adds source-type judgment, perspective analysis, corroboration, and confidence.
It is distinct from Data Integrity Preservation. Data integrity asks whether records remain accurate and consistent through their lifecycle. Source Provenance Triangulation asks what claims those records can support.
It is distinct from Source-of-Truth Assignment. Source-of-truth assignment chooses an authoritative representation for operational consistency. Source Provenance Triangulation may conclude that no source is fully authoritative and that confidence must vary by claim.
It is distinct from Epistemic Inclusion Design. Epistemic inclusion addresses who gets to contribute to knowledge-making. This archetype addresses how sources, including included voices, support claims.
It is distinct from Narrative Construction Audit. Narrative audit examines event selection, causal sequencing, focal actors, and omissions. Source Provenance Triangulation examines whether the sources behind that narrative warrant the claims being made.
Variants and Near Names¶
Primary Source Proximity Check is a variant focused on ranking and qualifying firsthand sources by closeness, access, mediation, and context. It prevents the primary-source label from doing too much work.
Secondary Interpretation Layer Review focuses on later synthesis, translation, commentary, citation, and interpretive transformation. It is useful when derivative consensus may be hiding a weak or circular origin.
Perspective Triangulation focuses on differently positioned observers and records. It compares what each standpoint could know, what it might distort, and which voices are missing.
Chain of Custody Validation focuses on evidence handling and transformation. It is especially important when records, artifacts, logs, or documents may have been altered, contaminated, selectively transmitted, or edited.
Near names include source criticism, evidence triangulation, provenance tracking, source reliability assessment, citation review, witness comparison, and audit-trail review. These should usually be treated as methods or variants under the parent rather than separate archetypes.
Cross-Domain Examples¶
In historical analysis, the archetype helps reconstruct events from diaries, official records, photographs, newspapers, later memoirs, and scholarship without treating them as interchangeable evidence.
In incident review, it compares telemetry, logs, chat records, customer reports, and interviews so the final postmortem does not simply adopt the most vivid or senior narrative.
In journalism, it checks official statements against documents, affected parties, independent experts, and contemporaneous records before publishing a claim.
In research synthesis, it prevents several papers or reports that cite the same original source from being mistaken for independent confirmation.
In institutional memory, it preserves lessons learned with evidence links, confidence labels, dissent records, and update triggers so future teams know what the organization actually learned and what remains uncertain.
Non-Examples¶
A bibliography is not Source Provenance Triangulation. It may list sources, but it does not necessarily evaluate source type, provenance, proximity, perspective, corroboration, or confidence.
An archive is not the archetype. It stores or preserves materials. The archetype begins when those materials are evaluated as evidence for claims.
A footnote is not the archetype. It points to a source, but it can hide weak support unless citation lineage and source relevance are checked.
A single source-of-truth database is not the archetype. It may be operationally authoritative, but triangulation is needed only when claims require evidentiary comparison.
A polished historical essay is not automatically the archetype. It may contain good scholarship, but the archetypal intervention is visible only when source provenance and claim confidence are made inspectable.