Grounding judgment in the record.
RISCON Labs’ research begins with records accumulated through sustained observation of people, organizations, and the decisions made within them. Each system organizes these materials into distinct analytical layers. When a question is posed, it retrieves relevant evidence from its higher-level analytical corpus to develop an analysis.
The central aim is to make the relationships between records, interpretation, and judgment open to scrutiny.
- Foundation
- A long-term research archive
- Method
- Layered, question-led extraction
- Standard
- Traceable grounds, explicit limits
An archive developed over time.
Our research rests on a personal research archive spanning the early 1990s to 2026. It brings together personal records, materials accumulated through professional practice, and subsequent reflection. Firsthand records of organizational management form one strand of this longer history of observation and inquiry.
The research value of these materials lies in the opportunity to examine how situations unfolded. We consider what was known at the time, which alternatives were considered, how decisions were reached, and what subsequently changed. Reading these elements together allows us to analyze the gaps between intent, agreement, and implementation.
- Contemporaneous records
- Records produced in the course of work and decision-making are examined in light of when and why they were created, and the circumstances in which they arose.
- Retrospective accounts
- Personal narratives and later reflections provide additional context but are kept distinct from records created at the time of the events.
- External research
- Publicly available research, comparative cases, and domain-specific knowledge supplement the basis for judgment, with sources and reference dates clearly identified.
The coverage and density of the record vary by period and subject. Individual records reflect their authors’ perspectives and the constraints under which they were produced. The existence of documentary evidence does not, in itself, establish that the record is complete or representative, or that a causal relationship exists.
Distinct analytical layers, grounded in original sources.
Geological strata offer an analogy for this structure. Original source materials form the foundational layer. Above them sit contextualized accounts of events and higher-level corpora assembled for analytical purposes. New questions guide the retrieval of relevant materials from these corpora. To check an interpretation, we return to the underlying context within the limits of permitted access.
Select a layer to read its role and review criteria.
Original source archive
- Analytical role
- We distinguish original source materials from subsequent interpretations. This foundational layer provides a point of reference for examining what later analyses rely on.
- Outputs
- This layer provides original records and verifiable information about their provenance: when, under what circumstances, and for what purpose they were created.
What facts can this record establish, and what falls outside its scope?
Events in context
- Analytical role
- Related records are brought together around a situation or decision process. We reconstruct the sequence of events and the information available at the time, while addressing gaps and conflicting accounts.
- Outputs
- This layer provides event-level accounts of the situation, options, actions, and subsequent developments, with links to supporting evidence.
Have we distinguished what was proposed, what was agreed upon, and what was actually done?
Higher-level analytical corpus
- Analytical role
- Events are organized by research question, field, and underlying mechanism. A corpus is a collection of materials assembled for analysis. We establish a basis for comparison while preserving the conditions specific to each case.
- Outputs
- This layer provides structured cases, thematic collections of materials, alternative interpretations, and links back to the underlying context.
What can be compared using the same criteria, and which differences might alter the interpretation?
Question-specific evidence
- Analytical role
- Relevant materials are retrieved from the higher-level corpus in response to each question. We assess their relevance, incorporate counterexamples and external research, and make explicit the assumptions required to move from evidence to interpretation.
- Outputs
- This layer sets out supporting and contrary evidence, unresolved questions, and the conditions under which an interpretation would need to be revised.
Does the retrieved material support the claim, or is it only superficially related to the question?
Human judgment
- Analytical role
- Readers consider the evidence in light of current circumstances, competing objectives, and the consequences of available choices. Judgment also involves identifying the changes that would warrant reconsideration.
- Outputs
- The resulting judgment makes its reference date, key assumptions, and conditions for review explicit.
Who is accountable for the judgment, and what new evidence would prompt a change in the decision?
This structure describes the research method shared across the projects. The materials actually made available, the functionality implemented, and the scope of validation results are specified separately for each project.
Higher-level analytical corpusDifferent questions require different evidence.
The same materials can give rise to different questions. A delay, for example, can be examined in terms of authority, information flows, incentives, or resource constraints. The evidence required depends on which explanation is being assessed.
- 01Identify judgment scope & valid bounds
Define the question
Specify the judgment under review, the relevant reference date, and the conditions under which it would hold.
- 02Corpus query & event retrieval
Retrieve relevant evidence
Select relevant events and comparative materials from the higher-level corpus.
- 03Rival hypotheses & falsification
Compare explanations
Examine the context, alternative explanations, and evidence that could overturn the claim.
- 04Traceable grounds & review criteria
Establish the basis for judgment
Set out the evidence supporting the judgment, the remaining uncertainties, and the conditions for reconsideration.
Did the approval process contribute to the delay?
- Observation
- The records examined show that a delay occurred after approval.
- Explanation to examine
- The approval process may have constrained implementation. The sequence of events alone, however, does not establish this explanation.
- Further evidence required
- The inquiry would examine information available before approval, implementation constraints, changes in external conditions, and comparable cases in which the same procedure was followed but no delay occurred.
The credibility of an interpretation rests on how evidence is handled.
Our shared methodology requires each analysis to make four distinctions explicit.
Documented facts and inference
Fact vs InferenceWe distinguish what is directly established by a source from the researcher’s interpretation and from material drawn from external knowledge. Cases constructed for explanatory purposes are identified as hypothetical or reconstructed.
Historical context and present-day application
Past vs PresentWe distinguish the time of the event from the time of analysis. We examine whether the conditions underlying an explanation of past events still hold today.
Repeated observation and generalization
Observation vs GeneralizationRepeated observations provide a basis for developing hypotheses. Generalization to other settings requires examination of comparative cases and counterexamples, as well as scrutiny of selection bias and gaps in the source material.
Research access and public disclosure
Access vs DisclosureThe scope of research use is considered separately from the scope of external disclosure. Where identifying information about individuals or organizations is unnecessary to the analysis, accounts focus instead on roles, relationships, and operating conditions.
A shared methodology, with field-specific questions.
RISCON Labs’ projects share a method for organizing records, examining context, and establishing the basis for judgment. Research questions, domain-specific knowledge, and criteria for evaluating answers are defined according to each project’s purpose.
A shared research foundation does not make findings in one field automatically valid in another. Each application requires an explicit determination of which materials are relevant, what can be compared, and what level of evidence is needed.
Methodological references
These works provide background on related research approaches. Their inclusion does not imply external validation or evaluation of RISCON Labs’ materials or systems.
- Ann Langley · 1999Strategies for Theorizing from Process Data
This article discusses analytical strategies for developing explanations from the unfolding sequence of events.
- W3C · 2013PROV-Overview
This document introduces a framework for describing the origins and derivation of information.
- Patrick Lewis et al. · 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
This paper presents an approach that combines retrieval from external knowledge sources with generation.