Witness / Source Comparison¶
Method — instantiates Source Provenance Triangulation
Compares firsthand accounts or records by role, access, incentive, timing, memory risk, and corroboration against non-testimonial evidence.
When several people saw the same event and tell it differently, the loudest or most senior account tends to win — which is exactly the wrong basis for belief. Witness / Source Comparison is the method that lays multiple firsthand accounts side by side and weighs each by what its author could actually have known and had reason to say: their vantage and access, their role and incentives, how long ago they observed, and how much their memory has had to reconstruct. Its defining move is comparative weighting of testimony: it does not accept or reject accounts one at a time, it ranks them against each other on proximity and standpoint, so a dim but well-placed observer can outweigh a vivid but distant or interested one. The output is a graded set of accounts, not a single chosen truth.
Example¶
A grinding machine severs a guard and injures an operator on a factory floor; four people give statements. The supervisor, who arrived after the noise, describes the operator "reaching into the running machine." A nearby operator, facing away, heard but did not see. The injured worker, still shaken, recalls the guard "already loose that morning." A maintenance tech who serviced the line the previous week describes the guard's latch as worn.
Witness / Source Comparison lays these out and weighs them. On proximity: the nearby operator heard but had no line of sight; the supervisor saw only the aftermath; the injured worker was closest but under acute stress and possible reconstructive bias; the tech saw the guard days earlier, not the event. On standpoint: the supervisor has an incentive to locate fault in operator behaviour rather than maintenance; the injured worker's memory of "already loose" may be shaped by hindsight. The method then weights the accounts against the non-testimonial record — the maintenance log and the machine's own condition — and finds the tech's worn-latch account and the physical evidence mutually reinforcing, while down-weighting the supervisor's aftermath-based inference. The report is a ranked, reasoned set of accounts that keeps the vivid supervisor narrative from silently becoming the finding.
How it works¶
- Profile each account's vantage. For every witness, record where they were, what they could see or hear, when they observed, and how mediated their access was — closeness is not only about time.
- Profile each account's standpoint. Note role, incentive, audience, and the reconstructive pressures on memory, so a motivated or stress-shaped account is located rather than trusted flat.
- Weight, don't average. Rank the accounts by proximity and standpoint; a well-placed disinterested observer outweighs a distant or interested one, and equal treatment is explicitly refused.
- Anchor against non-testimonial evidence. Check the weighted testimony against records and physical traces that don't share the witnesses' vantage or incentives.
Tuning parameters¶
- Proximity model — how you score closeness (time, line of sight, role, mediation). A richer model separates the truly well-placed witness from the merely nearby, at the cost of harder judgment calls.
- Memory-decay weighting — how heavily you discount accounts for elapsed time and reconstruction risk. Aggressive discounting protects against false memory but can bury a genuine late recollection.
- Incentive sensitivity — how strongly a conflict of interest lowers an account's weight. High sensitivity resists self-serving narratives but risks unfairly discounting an interested but accurate witness.
- Anchor strictness — how much weight the non-testimonial record is allowed to override testimony. Strong anchoring guards against consensus illusion; too strong ignores what only witnesses could know.
When it helps, and when it misleads¶
Its strength is that it stops the most authoritative or most vivid account from winning by default, and gives a principled reason to prefer a modest, well-placed witness over a confident, distant one. It is the core discipline of accident investigation, oral history, and any postmortem drawing on human accounts.
Its central failure mode is the fragility of memory itself: confident testimony can be sincerely, elaborately wrong, because post-event information reshapes recollection — the misinformation effect documented in eyewitness research.[1] The classic misuse is credibility-by-demeanor — weighting an account by how assured or senior the witness sounds rather than by their access and incentives, which is precisely the bias the method exists to correct. The discipline that keeps it honest is to weight on vantage and standpoint, never on confidence or authority, and to hold the ranked testimony against non-testimonial evidence rather than letting the most persuasive human account close the question.
How it implements the components¶
event_proximity— scores each account by closeness to the event: line of sight, timing, role, and mediation, not merely temporal nearness.perspective_bias_profile— locates each witness by role, incentive, audience, and memory pressure, so an account is interpreted from its standpoint rather than trusted flat.reliability_weighting_rule— ranks the accounts unequally by proximity and standpoint, refusing to average testimony that deserves different weight.
It weighs accounts against each other, so it does not trace whether repeated documentary sources share one origin or map inherited framing (corroboration_check, interpretive_layer_map — Citation Lineage Review); nor does it reconstruct a single source's authenticity and issue an admissions verdict (provenance_record, confidence_label — Source Criticism Protocol, its nearest twin, which gates one source in isolation where this method ranks many against each other).
Related¶
- Instantiates: Source Provenance Triangulation — supplies the comparative weighting of firsthand accounts that keeps the most vivid or senior narrative from becoming the finding.
- Consumes: Audit Trail Review — supplies the non-testimonial records this method anchors weighted testimony against.
- Sibling mechanisms: Audit Trail Review · Citation Lineage Review · Source Criticism Protocol · Conflicting Source Table · Triangulation Matrix · Chain-of-Custody Record
Editorial Notes¶
Form Classification¶
Form family: Assessment, Review & Assurance
Rationale: Witness / Source Comparison operates as a bounded evaluation of existing evidence or work that produces a finding or disposition because it compares firsthand accounts or records by role, access, incentive, timing, memory risk, and corroboration against non-testimonial evidence.
Independent corroboration: The frozen evidence defines Witness / Source Comparison as 'Compares firsthand accounts or records by role, access, incentive, timing, memory risk, and corroboration against non-testimonial evidence', so its operative form is Assessment, Review & Assurance.
Nearest alternative: Analysis, Modeling & Optimization — Witness / Source Comparison includes features of an analytical, modeling, inference, comparison, or optimization procedure that derives insight or a solution, but its defining operation is a bounded evaluation of existing evidence or work that produces a finding or disposition.
Review outcome: Independent reviewer agreement; medium confidence.
Origin Attribution¶
Primary origin: History & Historiography
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Comparing firsthand accounts by access, timing, incentive, memory risk, and corroborating records is historical source criticism. National Archives guidance centers provenance, context, perspective, and cross-source corroboration when evaluating primary evidence; courts and intelligence apply parallel credibility tests.
Related originating lineages:
- Accounting & Auditing — Accounting and audit's ledger, variance, evidence, and assurance tradition supplies an independent formative lineage for the mechanism's witness source comparison logic.
- Criminology & Forensic Studies — Criminology, forensic evidence, and chain-of-custody practice has a distinct contributing or parallel lineage for the mechanism's defining operation: compares firsthand accounts or records by role, access, incentive, timing, memory risk, and corroboration against non-testimonial evidence.
- Ethnography & Qualitative Methods — Ethnography and qualitative comparative inquiry has a distinct contributing or parallel lineage for the mechanism's defining operation: compares firsthand accounts or records by role, access, incentive, timing, memory risk, and corroboration against non-testimonial evidence.
- Law & Governance — Legal doctrine, regulatory governance, and procedural accountability has a distinct contributing or parallel lineage for the mechanism's defining operation: compares firsthand accounts or records by role, access, incentive, timing, memory risk, and corroboration against non-testimonial evidence.
- Security Studies & Intelligence Analysis — security_intelligence contributes security engineering, threat analysis, and intelligence practice to this mechanism's defining operation—Compares firsthand accounts or records by role, access, incentive, timing, memory risk, and corroboration against non-testimonial evidence—without displacing the selected primary historical lineage.
Review resolution: The blind reviewers disagree on primary lineage (law_governance versus history_historiography). Authoritative or primary research supports history_historiography as the best historical origin: Comparing firsthand accounts by access, timing, incentive, memory risk, and corroborating records is historical source criticism. National Archives guidance centers provenance, context, perspective, and cross-source corroboration when evaluating primary evidence; courts and intelligence apply parallel credibility tests. The cited U.S. National Archives, Using Primary Sources and Assessing Historical Evidence directly supports the mechanism's defining operation. All independently supported contributing domains are retained without an arbitrary cap. origin_mode=cross_disciplinary_synthesis records lineage, while domain_reach=multi_domain records later applicability separately from provenance.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Researched adjudication after independent review; high confidence.
Sources consulted:
References¶
[1] Loftus, Palmer. "Reconstruction of Automobile Destruction: An Example of the Interaction Between Language and Memory". Journal of Verbal Learning and Verbal Behavior 13(5): 585–589 (1974). Demonstrates that post-event question wording can alter eyewitness estimates and later recollection. registry ↩