Sustainable Catalyst Narrative Evidence Environment
Narrative Risk
Narrative Risk is a connected claim, evidence, timeline, and review system
for evaluating stories that influence public understanding, institutional action,
research interpretation, strategy, policy, and decision-making.
It separates what is being asserted from what is supported, preserves the source
and evidence chain behind each claim, tracks contradiction and narrative change,
records uncertainty and consequence, and produces reviewable artifacts
instead of opaque reputation scores or automatic truth judgments.
a narrative assessment should preserve the claim, source, evidence,
timeline, uncertainty, contradiction, review state, and decision context.
It should never hide weak evidence behind a confident label.
System overview
From isolated claim scoring to connected narrative evidence review
Narrative Risk now serves as the platform’s structured layer for contested claims,
evidence quality, changing narratives, uncertainty, interpretation risk,
consequence, review decisions, corrections, and downstream handoffs.
Capture
Record the claim precisely
Preserve the exact assertion, claimant, audience, implied conclusion,
affected decision, topic, geography, and publication context.
Ground
Connect the claim to evidence
Attach sources, snapshots, quotations, datasets, methods,
corroboration, contradiction, limitations, and missing evidence.
Track
Follow narrative change over time
Record when claims emerge, spread, strengthen, weaken,
reverse, fragment, become outdated, or are superseded.
Review
Preserve judgment and next action
Document confidence, uncertainty, risk factors,
consequence, review status, decision notes, corrections, and escalation.
Live workspace
Evaluate a claim in the browser
The public demo evaluates source type, evidence strength, uncertainty,
narrative volatility, stakeholder pressure, time sensitivity,
consequence, review status, and source count.
It produces a risk score, level, flags, review actions,
decision note, and structured JSON export.
do not enter confidential investigations, private allegations,
legal strategy, sensitive personal information, protected sources,
client records, or non-public intelligence.
The output is educational and exploratory and requires human review.
Interactive demo
Catalyst Narrative Risk
Build a traceable review record linking claims to sources, evidence excerpts, narrative structure, assumptions, wording variants, uncertainty, and human judgment.
View JSON export
{}
This demo is educational and exploratory. It structures review; it does not verify truth, provide legal advice, approve communications, or replace professional judgment.
Current capabilities
Eight components of a reviewable narrative record
The system preserves the structure required to inspect how a claim,
story, interpretation, or public position was assessed.
01 · Claims
Canonical claim records
Store the assertion, claimant, claim type, affected audience,
implied action, context, scope, status, and related claims.
02 · Sources
Source records and snapshots
Preserve source identity, publisher, author, date, URL,
retrieval context, archived content, quotation, license, and access status.
03 · Evidence
Support, contradiction, and relevance
Distinguish direct, indirect, contextual, corroborating,
contradictory, outdated, disputed, absent, or methodologically weak evidence.
04 · Timeline
Narrative events and change
Track origin, amplification, revision, reversal,
correction, retraction, supersession, recurrence, and unresolved change.
05 · Assessment
Transparent risk factors
Evaluate source quality, evidence strength, uncertainty,
volatility, stakeholder pressure, time sensitivity, consequence, and diversity.
06 · Review
Human judgment and status
Preserve reviewer, rationale, confidence, limitations,
escalation, dissent, approval, rejection, update, and re-review requirements.
07 · Decisions
Decision and communication notes
Record what action was taken, which claim was accepted or bounded,
which caveats were retained, and what evidence would change the decision.
08 · Corrections
Revision and supersession history
Preserve corrections, clarifications, withdrawn statements,
later evidence, superseded assessments, and public-safe update records.
Claim and evidence model
Separate the assertion, source, evidence, interpretation, and decision
Narrative Risk prevents a score, label, headline, or summary
from becoming detached from the records that justify—or fail to justify—it.
Assertion
Claim
The exact proposition being evaluated,
including scope, qualifiers, implied conclusion, and affected decision.
Origin
Source and claimant
Who made the claim, where it appeared,
when it was published, and what access or archival context applies.
Record
Evidence item
The document, dataset, quotation, observation,
method, testimony, analysis, or reference used in the assessment.
Relationship
Support or contradiction
How strongly the evidence supports, limits,
contextualizes, conflicts with, or fails to address the claim.
Change
Timeline event
A dated emergence, revision, amplification,
correction, reversal, retraction, or supersession event.
Context
Narrative frame
The language, assumptions, stakeholder interests,
emotional framing, institutional context, and interpretation boundary.
Evaluation
Risk assessment
The declared inputs, weights, flags,
uncertainty, consequence, confidence, level, and model limitations.
Judgment
Review decision
The reviewer’s conclusion, rationale,
dissent, escalation, publication status, and next evidence requirement.
Use
Downstream handoff
The research, analysis, decision, publication,
support, communication, or institutional artifact that consumes the assessment.
Narrative review workflow
From contested claim to bounded decision record
Each step preserves the information needed to understand
why a claim was accepted, limited, rejected, escalated, corrected, or left unresolved.
- 01Capture
Record the precise claim, claimant, context, scope, audience, and implied action.
- 02Source
Attach primary and secondary sources, snapshots, quotations, data, and retrieval context.
- 03Compare
Identify corroboration, contradiction, source dependence, gaps, and outdated evidence.
- 04Contextualize
Map timeline, framing, stakeholder pressure, volatility, and institutional context.
- 05Assess
Evaluate evidence strength, uncertainty, consequence, time sensitivity, and source diversity.
- 06Decide
Record confidence, decision note, caveats, dissent, escalation, and next evidence needs.
- 07Correct
Update, clarify, supersede, withdraw, or reopen the assessment as evidence changes.
- 08Publish
Create internal, partner, or public-safe artifacts with visible boundaries and provenance.
Applications
Narrative discipline across research, policy, products, and institutions
Narrative Risk is most useful where claims travel quickly,
sources vary in quality, interpretation matters,
and later review must reconstruct what was known at the time.
Research
Contested findings and evidence synthesis
Separate reported findings, methods, interpretations,
limitations, disagreement, replication status, and later evidence.
Public intelligence
Changing events and source uncertainty
Review event narratives, official statements,
public reports, source conflicts, freshness, and unresolved claims.
Impact
Sustainability and public-interest claims
Keep impact, climate, development, governance,
technology, and institutional claims aligned with actual evidence.
Institution
Communications and decision records
Preserve the reasoning behind public language,
strategic positions, corrections, disclosures, and institutional responses.
Platform connections
How Narrative Risk connects to Sustainable Catalyst
Narrative Risk remains the specialized claim-review product
while exchanging sources, evidence, data, analysis,
experiments, decisions, and publication records
with the wider platform.
Canonical records
Catalyst Data
Receive entities, periods, observations, measurements,
sources, provenance, validation states, and evidence-chain records.
Sources and publication
Knowledge Library
Connect claims to documents, source snapshots, citations,
quotations, collections, relationships, corrections, and durable records.
Research routing
Research Librarian
Route missing evidence, related sources, conflicting records,
research questions, title discovery, and product-specific follow-up.
Public context
Site Intelligence
Review country, event, indicator, public-source,
geospatial, temporal, humanitarian, economic, and institutional narratives.
Quantitative interpretation
Workbench and Catalyst Analytics R
Test calculations, assumptions, indicators,
scenarios, uncertainty, sensitivity, model limits, and quantitative claims.
Empirical validation
Research Lab
Connect claims to experiments, datasets, observations,
calibration, validation, reproducibility, and scientific provenance.
Strategic framing
Catalyst Canvas
Evaluate problem statements, assumptions, stakeholder narratives,
hypotheses, experiment rationales, and decision-readiness claims.
Decision governance
Decision Studio
Send claim assessments, uncertainty, evidence gaps,
contradictions, consequence, confidence, and review decisions into Decision Packets.
Shared infrastructure
Platform Core
Register claim IDs, source snapshots, evidence records,
provenance activities, review decisions, typed artifacts,
compatibility, access boundaries, and trust metadata.
Outputs and interoperability
Portable claim-review records instead of isolated scores
Narrative Risk can produce machine-readable and human-readable artifacts
that preserve the assessment inputs, evidence chain,
timeline, rationale, limitations, and review state.
Assessment
Narrative-risk report
Claim, score inputs, flags, level, evidence summary,
uncertainty, consequence, decision note, and review actions.
Evidence
Claim-evidence matrix
Sources, quotations, snapshots, support,
contradiction, relevance, limitations, provenance, and evidence gaps.
History
Timeline and correction log
Emergence, amplification, revision, reversal,
correction, retraction, supersession, and current status.
Exchange
Public-safe and connected artifacts
JSON, review briefs, decision handoffs,
publication notes, evidence manifests, and typed platform packages.
Governance and integrity
Review authority, provenance, and correction remain explicit
Narrative Risk should make disagreement and uncertainty inspectable,
not erase them through an apparently objective score.
Provenance
Evidence should point backward
Preserve source identity, snapshot, retrieval,
quotation, transformation, reviewer, date, and downstream use.
Dissent
Disagreement should remain visible
Record competing interpretations, reviewer disagreement,
alternative explanations, unresolved evidence, and minority positions.
Correction
Assessments remain open to revision
New evidence may clarify, weaken, reverse,
supersede, or reopen an earlier assessment or public statement.
Authority
Human review remains decisive
Scores, labels, summaries, rankings,
and AI-assisted analysis remain advisory and require accountable judgment.
Boundaries
Evidence discipline, not prediction, surveillance, or automatic truth
Narrative Risk structures claims and evidence.
It does not eliminate uncertainty, determine intent,
forecast events, establish liability, or replace domain expertise.
No prediction
It does not forecast events or markets
Narrative change may inform review,
but the system does not predict policy outcomes,
public reaction, financial performance, or future events.
No automatic truth
A score cannot determine factual correctness
Truth assessment still depends on source quality,
methods, context, expertise, corroboration, and accountable review.
No reputation laundering
Weak evidence must remain weak
The framework should not be used to make unsupported claims
appear verified, balanced, responsible, or institutionally approved.
No intent inference
Claims do not prove motive
Evidence about language or behavior does not automatically
establish deception, malice, negligence, coordination, or institutional intent.
No unrestricted investigation
Privacy and legal boundaries still apply
Personal data, private communications, protected sources,
confidential records, regulated material, and legal holds require proper controls.
No professional substitution
Qualified review may be required
Legal, journalistic, scientific, financial,
security, medical, regulatory, and investigative uses require appropriate expertise.
Next step
Make the claim, evidence, and decision path reviewable
Begin with the exact claim, preserve the source and supporting record,
compare contradiction and corroboration, map the timeline,
state uncertainty and consequence, and document the final review decision.