Provenance Tagging Protocol¶

Annotation protocol — instantiates Inclusive Membership Union Design

Attaches to each union member a durable record of which sources supplied it — with version, timestamp, and membership criterion — keyed to its canonical identity.

Version: v1 · 2026-08-24 · History
Mechanism #: 6836
Type: Annotation Protocol
Form family: Record, Log & Register
Solution family: Classification & Taxonomy
Problem family: Identity, Provenance & Integrity Failure
Problem subfamily: Collision, Membership & Feature Binding
Origin domain: Library & Information Science
Also from: Computer Science & Software Engineering
Instantiates: Inclusive Membership Union Design

Inclusion answers whether a member is in the union; provenance answers why. Provenance Tagging Protocol is the discipline of attaching, to every member, a standing record of its origins: which source collections contributed it, at what version and timestamp, and under what membership criterion each source applied. Its defining move is that the tag is keyed to the canonical member identity and travels with the member for its whole life — so at any later moment, for any member, you can answer "which sources put this here, and when?" This is not the momentary fold-on-collapse that happens while a union is built; it is the ongoing convention for how origin is recorded, structured, and preserved, so provenance survives re-runs, exports, and downstream copies rather than being reconstructed after the fact.

Example¶

A research team running a systematic literature review pulls candidate studies from PubMed, Scopus, and Embase and unions them into one screening set. Two years later a reviewer asks why a particular 2019 trial is in the corpus and whether it would still be retrieved today. Because the provenance protocol tagged each article at retrieval, the answer is on the record: this study carries sources: [pubmed@2023-04-01(query v3), scopus@2023-04-02(query v3)] keyed to its canonical DOI-based identity. It was not in Embase. The tag names the exact query version each database ran, so the team can reproduce the search, defend the corpus to a journal, and tell which studies would vanish if a database were dropped.

Without the protocol, the union would be a flat list of DOIs with no memory of how each got there — irreproducible, and impossible to audit when a reviewer challenges an inclusion. With it, every member carries its own explanation, addressable by its canonical identity.

How it works¶

Key the tag to canonical identity. Provenance is recorded against the member's canonical key, so it stays attached even after duplicates are collapsed and copies are made.
Capture the four facts. For each contributing source: which source, which version/snapshot, when it was read, and under what criterion the source counted the member.
Record contribution, not just presence. The tag lists every source that supplied the member (not merely the first), so corroboration structure is recoverable later.
Make it durable and structured. Store provenance in a fixed schema that survives export and re-computation, rather than as free-text notes that rot.

Tuning parameters¶

Provenance granularity — source-only, versus source-plus-version-plus-timestamp, versus full per-source membership criterion. Richer tags make the union fully auditable and reproducible but cost storage and capture discipline at ingest.
Retention horizon — how long historical provenance snapshots are kept. Long retention supports reproducing old unions; short retention saves space but loses the ability to answer "why was this here last year?".
Schema strictness — fixed controlled fields versus free-form annotation. Strict schemas are queryable and machine-checkable; loose ones are easier to write but degrade into unsearchable prose.
Capture point — tagging at first ingest versus at union build. Tagging early records the truest origin; tagging late is simpler but can miss a source that dropped out before the build ran.

When it helps, and when it misleads¶

Its strength is explainability and reproducibility: a union whose members each carry their origins can be audited, defended, reproduced, and pruned by source — none of which a bare merged list supports. The formal reference model here is data provenance as codified in the W3C PROV standard, which structures the "who/what/when" of how a data item came to be.^[n1]

Its failure mode is that provenance is only as honest as its capture. Tag late, tag partially, or record only the first contributing source, and the union grows an authoritative-looking but false origin record — which is worse than none, because people trust it. The classic misuse is treating provenance as a validity signal: a rich tag proves where a member came from, not that the member is correct, yet "well-documented" is easily misread as "verified." A further trap is letting provenance drift out of sync with the canonical key after a re-resolution, so tags point at a member that no longer exists. The guarding discipline is to capture provenance at ingest, in a strict schema, keyed to canonical identity, and to re-key tags whenever identity resolution changes — treating the tag as a first-class part of the member, not a decoration.

How it implements the components¶

provenance_retention_record — its core: the durable, structured record of every source, version, timestamp, and criterion behind each member.
canonical_member_identity_key — used as the anchor the provenance is filed under, so origin data stays bound to the member across collapse and export.

This protocol records origin as a standing convention; it does not itself choose which record survives when duplicates collapse (duplicate_collapse_policy) — that merge decision, and the one-time fold of provenance during it, is Deduplicating Union Pass, the nearest twin: this protocol defines how provenance is structured and kept; the pass merely folds it at collapse time.

Instantiates: Inclusive Membership Union Design — supplies the provenance record that makes the union auditable and reproducible.
Consumes: Canonical Identity Resolution Pass supplies the canonical key each provenance tag is filed under.
Sibling mechanisms: Canonical Identity Resolution Pass · Deduplicating Union Pass · Inclusive-OR Membership Test · Overlap and Coverage Dashboard · Type Compatibility Checklist · Union Delta Review · Union Specification Sheet

Editorial Notes¶

Form Classification¶

Form family: Record, Log & Register

Rationale: Provenance Tagging Protocol operates as a persistent ledger, log, register, or case record that preserves history and traceability because it attaches to each union member a durable record of which sources supplied it — with version, timestamp, and membership criterion — keyed to its canonical identity.

Independent corroboration: The frozen evidence defines Provenance Tagging Protocol as 'Attaches to each union member a durable record of which sources supplied it — with version, timestamp, and membership criterion — keyed to its canonical identity', so its operative form is Record, Log & Register.

Review outcome: Independent reviewer agreement; high confidence.

Origin Attribution¶

Primary origin: Library & Information Science

Origin pattern: Cross-disciplinary synthesis

Present-day reach: Multi-domain

Rationale: Persistent source, version, timestamp, and membership metadata are rooted in cataloging, archival provenance, and information stewardship.

Related originating lineages:

Computer Science & Software Engineering — Data lineage systems materially shaped canonical identifiers and machine-preserved provenance tags.

Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.