Research evidence · change classification

New work, new version, or newly indexed?

A changing search result or registry count does not always mean the research corpus changed. The event must be classified before any public claim is updated.

Status: Evidence-backed publication-record guide · Current registry snapshot · No new canonical work or DOI version detected · Citation claims withheld · No change to peer-review, employment, advisory, production, or institutional-authority status

Four events that look similar in a search result

Observed eventWhat it establishesWhat it does not establishPublic action
New workA distinct scholarly object supported by a primary publisher or registry record and reconciled against title, authors, identifier, and lineage.Peer review, impact, deployment, employment, or advisory authority unless separately evidenced.Add one canonical work with its exact status and source.
New versionA new public state of an existing work, normally connected by a versioned DOI or publisher lineage.A second unique work or a change from preprint to peer-reviewed article.Update the current version while preserving earlier versions as provenance.
Newly indexed recordA registry has discovered or propagated metadata for an already known work or version.A new work, new version, independent peer review, or a new citation.Record index coverage without increasing the deduplicated work count.
New search visibilityA search engine now surfaces a page, profile, post, or registry record.That the snippet is accurate, primary, current, or sufficient for a substantive claim.Use the result for discovery, then verify the underlying primary record.

Two version lineages, not four works

The Reinforcement Learning Contour has Version 1 and Version 2 DOI records. Bounded Recursive Self-Improvement Under Human Governance also has Version 1 and Version 2 DOI records. In each case, the version pair belongs to one canonical work lineage. Both current publisher pages identify Version 2 as the latest version and state that it is not peer-reviewed. The first is conceptual research; the second reports a bounded local empirical experiment. Those evidence classes remain distinct.

A registry arrival can be late without being new research

Europe PMC record PPR1299882 is the index record for the already known Version 2 bounded-recursive-self-improvement preprint. Its appearance in Europe PMC changed registry coverage, not the underlying work, version, peer-review status, or claim scope. This is why the public catalogue records both source provenance and a separate deduplicated work count.

Current reconciled snapshot

On 29 September 2026, Crossref returned 15 exact-name DOI records, Europe PMC returned 14 author-query records, and PubMed returned four exact-author records. The canonical library remains at 13 deduplicated works. No exact-name Crossref record was created after the prior 25 September check, and no later version appeared on either checked Preprints.org manuscript page.

These totals count different units: DOI deposits, index records, a biomedical subset, and reconciled work lineages. None is a citation count. Citation counters and platform metrics remain outside the public claim set because they have not been normalized into the canonical claim ledger.

Claim boundaries

Peer-reviewed journal articles, conference work, preprints, conceptual research, the bounded empirical preprint, software, the locally evidenced EWCC production system, employment, and advisory work remain separate evidence classes. A registry or search delta cannot transfer evidence from one class to another. This guide makes no claim of clinical deployment, regulatory approval, production validation, present employment, advisory engagement, institutional representation, or citation impact.

Primary and canonical records

Evidence checked 29 September 2026 (Asia/Jerusalem). Search results were used for discovery only; identity, identifier, version, publication-status, and indexing statements were checked against primary publisher and registry records.