---
name: artbull-evidence-and-sources
description: Use when assembling and handling the non-visual evidence for an Art Bulletin article — objects, archives, primary documents, provenance, and technical findings — and citing them rigorously. Art-historical claims must rest on verifiable sources, not assertion. Guides evidence handling; it does not fabricate documents or attributions.
---

# Evidence & Sources (artbull-evidence-and-sources)

Beyond the object itself, an Art Bulletin argument rests on **documentary and material evidence**:
archives, inventories, contracts, letters, sale records, provenance chains, and technical/conservation
findings. This skill governs how you gather, weigh, and cite that evidence so an expert reviewer
trusts it. The looking lives in `artbull-visual-analysis`; here you handle everything else.

## When to trigger

- Building the documentary basis for dating, attribution, patronage, or reception
- Reconstructing a provenance or ownership history
- Weighing conflicting documents or an uncertain attribution
- A reviewer questioned a source, a transcription, or a chain of provenance

## Handling the evidence

1. **Go to primary sources.** Archival documents, inventories, contracts, wills, correspondence,
   sale and exhibition records — cite the **repository, collection, and shelf/inventory mark** so
   another scholar can find them.
2. **Quote and translate carefully.** Give original-language quotations where the wording matters,
   with accurate translations; note editorial interventions in transcriptions.
3. **Build provenance rigorously.** Document the ownership chain with sources for each link; flag
   gaps honestly, and surface any **spoliation / restitution** concerns rather than burying them.
4. **Weigh attribution evidence.** Combine documentary, visual, and technical grounds; state the
   degree of certainty ("attributed to," "circle of," "documented as") and what would change it.
5. **Use technical art history where it bears.** X-radiography, infrared reflectography, pigment and
   support analysis, dendrochronology — cite the report and who performed it; do not overstate what
   science proves.
6. **Distinguish evidence from inference.** Separate what a document *says* from what you *infer*;
   keep silent documents from being made to speak.

## Anti-patterns

- Asserting a date or attribution with no documentary or technical basis
- Citing a source you have not seen (or a digital surrogate as if it were the original) without saying so
- Provenance with unexplained gaps presented as continuous
- Over-claiming what a scientific report establishes
- Burying a restitution / authenticity problem the evidence raises

## Source-strength ledger an Art Bulletin referee will apply

Specialist referees for the College Art Association's quarterly weigh the *kind* and *traceability* of
evidence, not its volume.

| Evidence type | Trusted when… | Referee distrusts when… | Strengthening move |
|---|---|---|---|
| Archival document | Repository, collection, and shelf-mark cited; consulted in person | Cited from a secondary source as if seen | Verify the original, or name the surrogate used |
| Provenance link | Each owner change has its own source | A continuous chain papering over a gap | Flag the gap; never present a break as continuity |
| Technical report | Method, analyst, and date named; claims stay within the science | "Science proves" overreach (IRR, XRF, dendro) | State what the report can and cannot establish |

## Worked vignette: reconstructing a contested provenance

Suppose an article dates a devotional diptych through its ownership history. A **1623 estate
inventory** in a named municipal archive (repository, fondo, and busta cited) records a "diptych with
the Virgin," the earliest secure link; the chain then jumps to an **1890s dealer ledger** and a **1931
museum accession file**, leaving a gap c. 1650–1890 the draft must flag rather than smooth over. A
**restitution concern** surfaces — the 1931 accession passed through a dealer active in a region of
known wartime spoliation — so the article notes a looted-art database check rather than burying it.
**Technical evidence** stays in its lane: dendrochronology of the support yields a date *consistent
with* the dating, reported as corroboration, not proof. The **certainty term** lands at "documented
from 1623, with a provenance gap," naming the document that would close it.

## Referee objections about sources, and the disciplinary fix

- *"You cite a document you have not seen."* Distinguish autopsy from secondhand citation; where you
  relied on a transcription or digital surrogate, say so in the note.
- *"A restitution / authenticity problem is buried."* Foreground spoliation or forgery concerns the
  evidence raises; candor about provenance ethics is a strength.

## Calibration anchors (hedge where uncertain)

- The apparatus carrying this evidence lives in **Chicago-style endnotes** (CMOS notes; confirm the
  current edition against the journal's current submission guidelines), documenting repositories,
  shelf-marks, and transcriptions in full — with original-language quotation and translation where
  wording bears weight for a global, multi-period readership.

## Output format

```
【Claim supported】dating / attribution / patronage / reception
【Primary sources】repository + collection + shelf-mark
【Provenance】chain with sources; gaps flagged
【Technical evidence】report + analyst (if any)
【Certainty】documented / attributed / circle-of, and what would change it
【Next】artbull-images-and-permissions
```

## Supplementary resources

- [`../../resources/external_tools.md`](../../resources/external_tools.md) — archives, provenance indexes, technical-art-history resources
- [`../../resources/official-source-map.md`](../../resources/official-source-map.md) — Chicago notes citation (CMOS ch. 14)