Gap
Missing evidence
No source has yet been found for a period, owner or transfer. A gap may matter, but it does not by itself prove that another account is false.
Conflicting provenance evidence exists when two or more sources appear to give incompatible accounts of a collectible’s identity, ownership, custody, location, acquisition, sale, manufacture, exhibition, publication or movement. The conflict may involve a date, name, serial number, photograph, label, invoice, catalogue entry, family account, certificate or legal record. It does not automatically prove fraud, but it does prevent the collector from treating every version as simultaneously established fact.
The disciplined response is not to select the most appealing story or the most prestigious source. It is to identify the exact proposition in dispute, preserve the evidence, test whether the disagreement is real, weigh each independent source for the claim it is being asked to prove, and state the conclusion at an appropriate level of confidence. A provenance file that keeps an unresolved contradiction visible is stronger than a polished chain built by suppressing inconvenient facts.
Collectors often use ‘conflict’ for every imperfect provenance record. That obscures the difference between an unanswered question, a source capable of several readings, a discrepancy that may be reconciled, and a claim that has been shown to be false. The classification affects both the language used and the amount of investigation required.
Gap
No source has yet been found for a period, owner or transfer. A gap may matter, but it does not by itself prove that another account is false.
More than one reading
The source supports several interpretations. A name may identify an owner, dealer, estate, family or unrelated person with the same initials.
Potentially reconcilable
Details do not align perfectly, but both sources may still be broadly correct—for example, sale date versus payment date or framed dimensions versus image dimensions.
Cannot both stand as understood
Two accounts make propositions that cannot both be correct without a different object, different event, changed physical state or mistaken source.
Shown to be false
Evidence establishes that the claim cannot be correct to a reasonable standard, such as a supposed 1910 label using a company name first adopted in 1934.
Intent requires proof
A document or story may be false without having been deliberately fabricated. Fraud should not be alleged merely because a claim is wrong or poorly supported.
A contradiction is material when it affects the conclusion a collector, buyer, insurer, researcher or institution is expected to rely on. The same discrepancy may be minor for one proposition and decisive for another.
The records may have been attached to the wrong object, variant, component or physical state.
An apparently continuous history may become impossible once dates, descriptions, materials or images are compared.
Two parties may have competing claims, or a supposed transfer may not have conveyed legal title.
An association with a maker, celebrity, military unit, event, production run or famous collection may be overstated.
The conflict may expose theft, unlawful export, coercion, illicit excavation, protected material or an improperly authorised disposal.
A price premium based on pedigree or historical use may not be justified if the supporting evidence is disputed.
Conflict can sit at several levels. Finding that level prevents unrelated discrepancies from being merged into one vague concern and helps identify the evidence most likely to resolve it.
Dimensions, inscriptions, serial numbers, edition, state, damage, accessories, frame, mount, packaging or component lists do not match.
Ask whether the records describe the same object, a changed object, a component, a duplicate or a similar example.
Manufacture, sale, ownership, inheritance, exhibition, shipment, export, restoration or publication dates appear incompatible.
First test whether the sources record different stages of the same event rather than one event on two dates.
A source confuses legal title with possession, consignment, storage, loan, agency, seizure, restoration or executor control.
‘With Dealer X’ and ‘from Collection Y’ do not automatically prove ownership by X or Y.
Initials, married names, titles, pseudonyms, translated names, partnerships, successor companies or estates are conflated.
Confirm address, dates, occupation, associates and collecting interests—not just a convenient name match.
The object appears to be in two places, or a location claim conflicts with transport, exhibition, wartime or export records.
Distinguish an owner’s address, storage site, temporary exhibition, branch office and catalogue-production location.
One source records a sale, another a gift, loan, exchange, inheritance, return, seizure, deposit or commission.
The wording may materially affect title, ethics, tax treatment and the interpretation of a forced or distressed transfer.
The item moves between ‘associated with’, ‘from the collection of’, ‘owned by’, ‘worn by’, ‘used by’ or ‘presented to’.
These are different propositions and often carry very different value premiums.
Sources disagree about title, export, restitution, voluntary sale, lawful excavation, deaccession, protected material or cultural-property status.
This is not merely a cataloguing problem; acquisition, sale or publication may need to pause.
Many conflicts begin with ordinary error: a misspelled name, reversed digits, incorrect lot number, misread handwriting, copied database field, wrongly attached photograph or inaccurate recollection. Institutional authority and professional presentation increase initial credibility, but neither guarantees that a record is correct or that it refers to the present object.
Other disagreements arise because sources were made for different purposes. Auction catalogues compress and promote; customs forms classify for regulation; invoices record accounting; insurance schedules record risk and value; family notes preserve memory; museum databases support retrieval. A short phrase may omit intervening owners, distinguish neither custody nor title, or repeat a claim supplied by the consignor.
Objects also change. They may be restored, reframed, rebound, trimmed, reassembled, separated from groups, fitted with replacement parts, relabelled or repackaged. Earlier and later sources may both be accurate while describing different physical states. Composite collectibles—medal groups, boxed sets, uniforms, fossils, manuscripts, framed works or assembled game sets—may contain components with different provenance histories.
Repetition is not corroboration when every later source derives from one earlier claim. A certificate, database entry, dealer biography and five sale listings may all trace back to one mistaken catalogue. Count independent evidential origins, not the number of documents repeating the wording.
Market language may strengthen by degrees: “possibly connected with” becomes “traditionally associated with”, then “from the collection of”, “owned by” and finally “personally used by”. Unless new evidence appears, the later certainty is not historical depth; it is claim inflation.
Family history may preserve genuine knowledge while compressing generations, dates and relationships. “Came through the family of X” can become “owned by X”; “acquired after the expedition” can become “brought back from the expedition”. Record oral evidence faithfully, then test it.
OCR mistakes, merged identities, image-to-record mismatches, lost punctuation, date-format changes and imported legacy errors can create digital confidence around a weak underlying record. Databases are research tools, not self-authenticating primary evidence.
Conflict becomes difficult to evaluate when observation, transcription, identification, interpretation and conclusion are compressed into one sentence. Keep the layers visible so another researcher can challenge one step without losing the underlying evidence.
Observation
A rectangular paper label is attached to the underside.
Transcription
The label reads: ‘J. H. Murray, 17 King Street.’
Identification
The name may refer to collector James Henry Murray.
Interpretation
The label may indicate ownership, dealership, repair, storage or inventory.
Bounded conclusion
The object was probably in Murray’s possession by approximately 1925.
There is no universal hierarchy in which official records always defeat photographs, or invoices always defeat oral testimony. A source can be strong for one fact and weak for another: a photograph may establish appearance and location but not legal title; an invoice may establish a transaction but identify the object only vaguely.
Does the source address the exact proposition in dispute? Evidence that someone owned a collection does not prove they owned every later-associated object.
How securely does it identify this object? Unique marks, serial numbers, inscriptions, photographs and distinctive damage usually carry more identifying force than generic wording.
Was the source created near the event, and by someone able to know? A direct participant can still be mistaken or self-interested, but distance must be considered.
Do several sources represent separate evidential origins, or do they all copy one catalogue, certificate, database entry or family note?
Is this the original, a scan, transcript, excerpt, quotation or paraphrase? Missing pages, cropped reverses and removed qualifiers can change meaning.
Does the source show alteration, overwritten dates, inconsistent fonts or ink, metadata anomalies, broken signatures, unexplained seals or missing attachments?
Does the source agree with itself, and with established business dates, biographies, production records, geography, archive holdings and transport possibilities?
An invoice, customs declaration, auction catalogue, memoir, certificate and family note were created for different purposes and may simplify, promote or protect different interests.
Who supplied the document or image, when was it first seen, can the original be located, and could it have been transferred from another object? Evidence has provenance too.
The method below is deliberately slower than choosing the most persuasive-looking document. It creates an auditable research trail, reduces confirmation bias and protects both the object and the future user of the record.
Replace ‘the provenance is inconsistent’ with a testable statement: ‘Source A identifies Mary Ellis as owner from 1934, while Source B states that the object remained with Thomas Green until 1938.’
Record the object and every source in its current state. Do not clean, remove, reposition, annotate or ‘correct’ evidence before it has been documented.
Do not merge conflicting accounts into a single narrative. Record who created each source, when, what event it describes, whether it is original or derivative, its exact wording, identifiers, limitations and relationship to other sources.
Place every dated event in order without choosing a preferred story. A timeline often reveals that an apparent contradiction is a loan, later acquisition, catalogue date, payment date or post-restoration description.
Retain original wording and add a justified normalised form. ‘Circa 1920’ remains approximate; ‘spring 1920’ does not become 1 March; ‘with X’ does not become ‘owned by X’.
Group copied and derivative sources into source families. Ten later references to one mistaken auction entry are one evidential line, not ten independent confirmations.
Ask whether the records concern different objects, components, editions, physical states, dates within one transaction, ownership interests, custody roles or similar examples.
List plausible explanations before choosing one: catalogue error, oral account attached to another item, loan rather than ownership, moved label, later alteration, duplicate example or deliberate fabrication.
Decide what could separate the hypotheses: stock books, photograph reverses, complete estate inventories, shipping records, payment ledgers, address histories, serial registers, conservation analysis or installation photographs.
Ask what would make the preferred claim impossible, what records should exist if it were true, and what alternative owner or object fits the evidence better.
If a certificate once treated as independent is shown to reproduce a 1987 catalogue entry, record why its evidential weight changed. Do not silently overwrite the earlier assessment.
Record whether the conflict is resolved, probably resolved, open, disproved, attributable to different objects, or suggestive of fabrication requiring specialist review.
A source affirmatively supports a proposition: an inventory lists the object by serial number, a signed receipt records a sale, or a dated photograph shows it in a known room. The source still needs to be assessed for identity, integrity and meaning.
An expected record is absent. That absence matters only if the record set is reasonably complete, covers the relevant period, would normally include the object and would identify it adequately. “Not found” is not the same as “never existed”.
A source expressly records the fact, such as a receipt stating that A sold the object to B. Direct evidence can still be false, incomplete or attached to the wrong object.
Several independent facts may combine persuasively: a stock book, buyer’s diary, later photograph, inventory label and estate list. No single source proves ownership, but their convergence may make the conclusion highly probable.
Do not turn an unresolved conflict into a binary verdict merely to make a catalogue entry look complete. Use controlled language that preserves what is known, the source of the claim and the reason for any qualification.
The object was recorded in the Ellis collection by 1947.
The combined inventory, photograph and correspondence make Ellis ownership highly probable.
The available evidence favours Ellis ownership, although the exact acquisition date remains uncertain.
The label may connect the object with the Ellis collection, but no independent confirmation has been found.
According to family tradition, the object belonged to Ellis.
Ellis ownership is stated in the 1984 sale catalogue but conflicts with a contemporary Green family inventory.
The conflict cannot presently be resolved from the available evidence.
The claim that Ellis acquired the object in 1920 is incompatible with evidence that it was manufactured in 1927.
The certificate contains chronological and material anomalies; its authenticity has not been established.
Overstated
Direct from the Langford estate.
Evidence-sensitive wording
Seller reported Langford estate provenance; current documentation confirms later purchase from the dealer, not direct estate transfer.
The later invoice supports a transaction but cannot prove the stronger estate-to-buyer chain asserted by the seller.
Overstated
Entirely nineteenth century.
Evidence-sensitive wording
Catalogued as nineteenth century in the cited sale; present examination suggests at least one later replacement component.
The catalogue remains part of the provenance file, but the physical conflict is disclosed rather than hidden.
Overstated
Owned by the family named in the note.
Evidence-sensitive wording
Family note associates the object with the named owner; an old label also links it to another collection. The relationship between the claims remains unresolved.
Neither account is selected merely because it is more attractive or easier to sell.
Not every discrepancy justifies abandoning an object. The response should be proportionate to what the conflict could change and how difficult the resulting harm would be to reverse.
Record and reconcile
Spelling variants, punctuation, translated titles, small measurement differences or approximate-date variation that do not affect identity or ownership.
Investigate before relying
Different acquisition dates, one uncertain intervening owner, unclear dealer role, inconsistent lot number or uncertain object variant.
Pause consequential decisions
Incompatible ownership chains, impossible chronology, conflicting serial numbers, a contemporary image contradicting location, or sale by someone who may not have held title.
Specialist escalation
Possible theft, unlawful export, coerced transfer, active ownership claim, falsified records, sanctions risk, human remains, sacred material, protected species or protected cultural property.
Evasive seller behaviour does not itself prove that the object is wrong, stolen or falsely documented. It does increase transaction risk because the buyer is being asked to accept a consequential uncertainty without reasonable access to the evidence needed to assess it.
A strong provenance system should allow several claims about the same event, preserve their sources and show how the assessment changed. Do not overwrite a disproved claim as though it never existed; retain its status and explanation so the research history remains auditable.
1936–1941: ownership uncertain. A 1968 dealer catalogue states that the object entered the Harold Wynn collection in 1936. A dated 1939 inventory in the Clara Voss archive appears to list an object with the same serial number. The dealer catalogue gives no source for the 1936 date. It remains possible that the inventory was copied incorrectly, that the object was on loan, or that one record refers to another example. Wynn ownership before 1941 is therefore unconfirmed.
A contradiction is not merely a defect to be eliminated. It may reveal a loan mistaken for a sale, an undisclosed dealer network, a family dispute, wartime movement, changing attribution, object alteration, the separation and reunion of components, or the gradual strengthening of a market narrative. The conflict may be the point at which the object’s real social and commercial history becomes visible.
Some conflicts will remain unresolved because records were destroyed, archives are restricted, witnesses have died, several identical examples existed, documents became separated from the object, or historical record-keeping excluded affected communities. Transparency is still possible even where certainty is not. The collector’s responsibility is to pass on a clearer, better evidenced and less distorted problem than the one they inherited.
Provenance research is not the manufacture of a complete story. When evidence conflicts, preserve the disagreement, investigate its cause, weigh each independent evidential line, document the reasoning and state no more than the surviving evidence supports. Honest uncertainty is not a weak provenance record; it is evidence of responsible stewardship.
Return to judging how much weight a source can safely carry for a particular provenance claim.
Return to the Provenance Evidence overview and its full topic list.
Continue to judging missing, incomplete and high-risk documentation across an ownership history.
Match the wording of a provenance conclusion to the strength and limits of the surviving evidence.
Understand why the person advancing a valuable or consequential claim must support it.
Separate an unexplained custody gap from a direct contradiction between two asserted histories.
Recognise when cautious association has grown into a stronger ownership or use claim without new evidence.