Purpose

The CSI Reference is intended to be a durable, independently verifiable scholarly resource. It documents the current conceptual architecture of Collective-State Inference while making clear which claims derive from established literature, which are interpretive syntheses, and which originate within the CSI research program.

Claim classification

  • Established scholarship: claims supported directly by peer-reviewed research, scholarly books, standards, or authoritative technical sources.
  • CSI synthesis: integrative interpretations that connect established literatures into the CSI framework.
  • CSI proposition: original theoretical claims that require future empirical evaluation.
  • Illustrative example: explanatory scenarios that clarify a concept but are not empirical findings.
  • Empirical finding: claims supported by reported data and identified as such.

Citation standard

Consequential theoretical, methodological, historical, and empirical claims should include a nearby citation. Each cited source must have a complete bibliographic record with a DOI, publisher page, or other stable identifier when one is available. Citations must support the specific sentence or passage to which they are attached; topical proximity is not sufficient.

Source selection

Priority is given to peer-reviewed journal articles, scholarly books from established academic presses, conference proceedings with formal review, standards, and primary technical documentation. Review articles may orient a topic, but foundational or disputed claims should be checked against primary sources. Sources are selected for evidentiary relevance rather than prestige alone.

Verification

Bibliographic details, identifiers, and claim-source relationships are checked before publication. Where sources disagree, the Reference should represent the disagreement rather than imply false consensus. Claims that cannot yet be verified are narrowed, labeled as propositions, or removed.

Versioning and corrections

Substantive changes to definitions, construct boundaries, evidence standards, citations, or interpretations are recorded in the Version History. Typographical and formatting corrections may be grouped as editorial changes. Confirmed errors are corrected promptly without silently rewriting the scholarly record.

Archived editions are immutable scholarly records. A later archival edition supplements rather than overwrites an earlier edition, preserving the ability to reconstruct what CSI claimed at each release.

Living Extension promotion and archival release

Living Extensions are the active research layer of the CSI Reference. They may be publicly cited and revised while their concepts, evidence, and boundaries continue to mature. Publication as a Living Extension does not guarantee inclusion in a later archival edition.

A Living Extension becomes eligible for archival promotion only when it clears four gates:

  1. Conceptual stability: the central position and construct boundaries are sufficiently stable that ordinary research developments are unlikely to reverse them.
  2. Scholarly grounding: relevant prior literature has been verified, the relationship to established scholarship is explicit, and the CSI novelty boundary is clear.
  3. System consistency: the extension is internally consistent with the CSI construct system, including its level-of-analysis, composition, evidence, validation, and uncertainty commitments where applicable.
  4. Meaningful scrutiny: the extension has survived substantive challenge through empirical work, adversarial literature review, peer-review feedback, theoretical stress-testing, or another documented form of scholarly scrutiny appropriate to the claim.

Promotion is milestone-driven rather than calendar-driven. At an Archival Release Gate, eligible Living Extensions are reviewed together for incorporation into the next archival edition. Completion of a consequential peer-review or empirical cycle may trigger such a gate, but does not automatically require promotion.

When an extension is promoted, its stabilized intellectual content is incorporated into a new archival edition. The prior edition remains unchanged and citable. Extensions that have not cleared the promotion gates remain Living Extensions and continue to develop independently of the archived record.

Research status

Reference articles may develop alongside unpublished research without depending on inaccessible material for their public explanations. Article metadata identifies the current status. Publication on this site does not by itself establish scholarly acceptance.

Generative AI assistance

Generative AI may support conceptual stress-testing, structural organization, language refinement, and iterative critique. The author remains responsible for verifying claims, references, interpretations, and final wording. AI-generated text is not treated as a scholarly source, and AI systems are not used as independent evidence for factual or theoretical claims.

Authorship and independence

The CSI Reference is maintained by Edward D. Clark as an independent researcher. Editorial responsibility, source selection, theoretical interpretation, and corrections remain with the author unless an article explicitly identifies additional contributors or reviewers.

Questions and corrections

Questions about sourcing, interpretation, or possible errors may be sent to corrections@collectivestateinference.org. Correction requests should identify the page, passage, and supporting evidence.