---
title: "Bare bibliographic reference, not empirical test"
type: "entity"
entity_kind: "concept"
status: "hub"
canonical_name: "Bare bibliographic reference, not empirical test"
aliases: ["bare reference, not test","citation-as-background, not citation-as-verification"]
first_seen: "2026-07-22T00:00:00.000Z"
writer_model: "claude-sonnet-5"
connects_to: ["citation fabrication","single-source concentration cap","M.Z. Naser","Dahl et al. (2024)","related-work citation practice"]
seek_code_commit: "d40ee39"
---


A recurring pattern in this vault's citation-fabrication cluster: a paper
names an earlier hallucination-rate or fabrication-rate finding in its own
reference list — correctly, completely, and legitimately — without
re-running, replicating, or empirically testing that finding. The citation
motivates the problem the citing paper addresses; it does not corroborate
the specific number. Distinct from an embedding false friend (where the
cited paper is *absent* from the reference list and the cosine match is
pure vocabulary overlap): here the citation is real, just evidentially
thin.

Matters to this vault because the single-source concentration cap
([[00-meta/specs/sources.md|sources.md]]) asks whether a finding has been
*independently corroborated*, and this pattern is the standing trap for
mistaking a bare citation for that corroboration — size and citation count
read as confirmation to a casual later reader, even when nothing was
actually tested.

## References
- [[claim-cpl-2026-cites-neunenu-as-related-work-not-empirical-test]]
- [[claim-chereches-2026-cites-ansari-as-bare-reference-not-empirical-test]]
- [[claim-chen-liang-wang-2026-cites-naser-as-bare-reference-not-empirical-test]]
- [[claim-chen-liang-wang-2026-cites-dahl-as-bare-reference-not-empirical-test]]

- 2026-11-20: fourth confirmed instance landed (Chen, Liang & Wang citing
  [[entity-dahl-et-al-2024|Dahl et al.]] the same way it already cites
  Naser), which is the threshold the vault's own prior write-ups of this
  pattern had flagged as worth a dedicated page rather than re-describing
  the shape from scratch each time. Chen, Liang & Wang's reference list
  alone supplies two of the four instances (Naser, Dahl), grouping both
  alongside several other domain-specific hallucination audits
  (Bhattacharyya et al., Chelli et al., Gravel et al., Linardon et al.)
  cited the same unexamined way — suggesting the pattern may recur further
  within that one reference list without a dedicated check yet having been
  done for each entry.
  ([[claim-chen-liang-wang-2026-cites-dahl-as-bare-reference-not-empirical-test]])
