Research Planning Strategy

Begin with the proposition, not the search box

The source is the thing in front of you: a label, note, invoice, photograph, catalogue entry, digital listing or memory. The claim is what someone wants that source to prove. Those are not the same. A dealer label may establish handling by a dealer without proving the earlier ownership printed beside it. A family note may show that a story was attached to the object by a particular date without proving the story itself.

Replace broad requests such as “find the provenance” with one or more answerable questions. Examples include whether a cited auction lot is this object, whether a gap between two owners can be bridged, when the object first entered the market, whether a label identifies an institution, whether the transfer was by sale, gift or inheritance, and whether dates or geography contradict the seller's account. Several questions may be necessary, but they should not be collapsed into one vague investigation.

Separate the propositions before you evaluate the evidence

Provenance enquiries often become confused because ownership, custody, association, attribution, authenticity, publication and legality are treated as interchangeable. They may support one another, but none automatically proves the rest. A photograph can establish association without ownership. An auction catalogue can establish that an object was offered without proving that it sold, that the consignor had good title or that the present object is the same example.

Establish the object's research identity before attaching a history

A convincing historical record is useless if it concerns another example. False provenance connections frequently begin when a researcher finds an object of the same type, title, maker or subject and assumes it is the present item. The plan must therefore begin with an identity profile detailed enough to discriminate this object from close comparators.

Preserve the supplied record before turning it into a story

Gather everything already associated with the object before searching externally: invoices, receipts, descriptions, correspondence, seller messages, export papers, family notes, insurance schedules, valuations, certificates, photographs, labels, digital listings, payment and shipping records, restoration records and previous research. Do not begin by rewriting these materials into a polished narrative. First preserve their original state and context.

Build the provisional timeline before drafting prose

A timeline exposes gaps, overlaps and undocumented transitions that fluent narrative can conceal. Use bounded language rather than invented precision: “by 14 May 1987”, “after 1972”, “between 1984 and 1989”, “documented in the collection by 1996” or “ownership not established”.

Investigate transitions, not merely the names on either side

A chain containing plausible owners can still be weak if none of the movements between them is evidenced. Each transition should be treated as a separate research problem, particularly where the object crosses a border, changes attribution, enters or leaves an institution, appears after a long absence, rises sharply in value or passes through a disputed period.

Not every blank period deserves equal effort. Prioritise gaps that affect identity, title, legality, ethics, cultural sensitivity, marketability or a consequential association claim. A routine undocumented household movement may be tolerable in a personal record; an unexplained wartime gap, alleged inheritance without estate evidence or recent appearance of an apparently ancient object may require immediate escalation.

Define the investigation before it expands indefinitely

Provenance research can continue without a natural end. Scope limits should control effort and expenditure, not predetermine the conclusion. Record what period, geography, evidence types and resources are included, and revise those boundaries consciously when discoveries justify it.

Make alternative explanations explicit before one becomes the story

A working hypothesis gives direction, but it must remain provisional. State what evidence should exist if it is correct and what would weaken or falsify it. This converts a preferred story into a testable proposition.

Map where evidence is likely to survive, then search in a rational order

Sources should be selected because they can answer the question, not because they are easiest to search. Build a source map across the object itself; transactional records; personal and family papers; institutional registers and files; publications; commercial and digital archives; legal and regulatory records; and knowledgeable communities. Community knowledge can reveal terminology and relationships, but it should remain attributed testimony until independently tested.

A sensible general search order

1. 01Preserve the object and all supplied documentation.
2. 02Extract names, dates, numbers, locations and distinctive terms.
3. 03Search the most discriminating object identifiers.
4. 04Verify known ownership or custody events.
5. 05Investigate the nearest unresolved transition.
6. 06Consult relevant published and digitised sources.
7. 07Approach institutions and archives with precise enquiries.
8. 08Interview witnesses where their knowledge is material and at risk of being lost.
9. 09Commission specialist, technical or legal work only where the likely benefit justifies it.
10. 10Reassess the question, hypothesis and source order after significant discoveries.

Work from known evidence and build the vocabulary the records may use

For many collectibles, research is most effective when it moves backwards from the current transfer: present collector, seller, auction, consignor, previous invoice, estate, dealer and earlier collection. This reduces the risk of finding an attractive historical record first and then forcing the object into it. Forward research can be equally valid when a strong early record exists, such as a factory dispatch register, studio inventory, excavation record, military issue ledger or institutional deaccession file. The two directions should meet; if they do not, the break remains a gap.

Ask archives and witnesses questions they can actually answer

“Do you have anything about this object?” transfers the entire research problem to the recipient. A useful enquiry identifies the object, the person or transaction, known references, date range, dimensions, photographs, suspected record series and the exact uncertainty to be resolved. Maintain a contact log with dates, questions, responses, restrictions, follow-up actions and citation or reproduction terms.

Oral-history planning requires similar discipline. Establish what is already documented, obtain consent, prepare open questions, avoid feeding the expected answer and ask witnesses to distinguish personal observation from family tradition or later inference. Interviewing witnesses separately may reveal whether apparent agreement is independent or merely repeated from one shared source.

Evaluate every source before allowing it to carry a claim

A genuine document may still be inaccurate, incomplete or promotional. A later source may preserve accurate information from an earlier record that no longer survives. Source criticism therefore asks who created the record, when, why, with what access, under what incentives, and whether it identifies this specific object.

Record what was searched and not found without overstating absence

A dated search log prevents repeated work, exposes the limits of the investigation and may narrow possible dates or demonstrate reasonable diligence. Record the repository, date, search terms, variants, filters, date range, records examined, result and access limitations.

Triage legal, ethical and cultural concerns early, not after the story is written

Research priorities change when an object may involve theft, forced sale, wartime seizure, conflict, colonial extraction, illicit excavation, human remains, sacred or ceremonial use, protected species, restricted export, national patrimony, restitution claims or private personal data. In these cases the question is not only how to strengthen provenance, but whether acquisition, retention, publication or sale is responsible and lawful.

Match the evidence standard to the purpose of the conclusion

The same material may be sufficient for a carefully qualified personal note and insufficient for a public catalogue, high-value sale, institutional acquisition or ownership dispute. Define the intended output before deciding how far the research must go.

Set review milestones, stop rules and claim-level confidence

Research should not end merely because the collector becomes tired, nor continue indefinitely because complete certainty has not appeared. Review the plan after the initial audit, after distinctive identifiers have been searched, after key archive responses, when the main hypothesis changes, when costs exceed likely benefit or when remaining inaccessible sources are unlikely to alter the conclusion materially.

Apply confidence to individual assertions rather than to the provenance as a whole. One history can contain a confirmed present owner, probable previous owner, uncertain transfer date and unsubstantiated earlier family story. The stated basis for a rating matters more than the label itself.

Preserve the full research trail, including paths that failed

The final record should allow another researcher to understand not only the conclusion but how it was reached. Preserve original files, unedited photographs, scans, transcripts, emails, archive references, catalogue details, URLs and access dates, search logs, permissions, restrictions, confidence assessments, superseded versions and rejected hypotheses. Never overwrite metadata-bearing originals with cropped, annotated or compressed derivatives.

Write only what the evidence can reasonably support

A finished provenance should distinguish documented fact from testimony and inference. “Acquired by Margaret Evans by 1978, when the object appears in her dated household inventory” is a documented statement. “Family testimony reports that she bought it in Edinburgh during the early 1970s, but no invoice has been located” preserves the oral account without converting it into a precise fact. “Ownership before Evans remains unknown” keeps the gap visible.

Recognise the strategic mistakes that create persuasive but weak provenance

A compact provenance research plan for a collection record

The plan does not need to become a formal academic document. It needs enough structure to stop claims, evidence, leads and decisions from collapsing into one narrative.

Twelve phases from securing evidence to preserving the conclusion