Matching Records to Objects

First decide what kind of match is being claimed

Many provenance errors begin before the comparison itself, because the proposed scope is left undefined. An exhibition number can identify one entry in one event. An auction lot number identifies a transaction grouping. A manufacturer's catalogue number may apply to thousands of examples. A catalogue raisonné number may identify a unique work, a design or an edition. The same-looking number can therefore perform very different jobs.

The collector should state the intended scope before judging the evidence. Is the claim that this exact object appeared? That one component formed part of a recorded set? That this edition or model was published? That a particular photograph has been reused? Or only that a previous seller asserted the association? The evidence needed rises with the specificity of the claim.

Build the match from the most discriminating evidence

Not all agreements carry equal weight. A reliable match normally combines several independent characteristics, beginning with features least likely to be shared by other plausible objects. Broad similarities help find candidates; object-specific features help eliminate them.

The hierarchy below is not a mechanical scoring system. A controlled identifier can be compelling, but only if its issuing system and attachment are trustworthy. A group of physical details can outweigh a doubtful loose label. The collector must judge both the distinctiveness of the feature and the reliability of the source in which it appears.

Read each record for what it can actually prove

Historical records were rarely created for future provenance researchers. An exhibition catalogue served visitors, a dealer list sold stock, an inventory controlled possessions, an auction catalogue described a lot, and an insurance schedule established values. The omissions and conventions of each source shape what can reasonably be inferred from it.

The published record may also be only the visible layer. Exhibition checklists, loan agreements, condition reports, transport records, installation photographs, annotated catalogues and internal inventories can contain stronger identifying evidence than the polished catalogue entry.

Match illustrated objects systematically

Historical photography can provide powerful object-specific evidence, but resemblance should be broken into testable features. Prioritise stable characteristics: inscriptions, manufacturing defects, print registration, grain or veining, enamel loss, stitching, plate tone, page damage, label residue and old repairs. A combination of persistent features is more persuasive than overall likeness.

Separate the principal object from its presentation. Frames, boxes, mounts, stands, cases and detachable accessories can change, while an old frame or box can later contain a different object. Ask three questions independently: does the object match, does the presentation match, and is there evidence that they remained together?

Allow for perspective, cropping, mirrored negatives, retouching, glare, halftone screening, colour shift, lens distortion and later restoration. Digital overlays can assist comparison, but resizing and warping can manufacture apparent alignment. An overlay is an analytical aid, not proof by itself.

Use convergent identification when no photograph survives

A text-only record can still support a strong match when several independent details create a highly discriminating profile. Maker, dimensions, a dated inscription, named lender, unusual damage, frame label and exhibition number may be individually non-unique but compelling in combination.

The decisive question is not whether the present object can satisfy the description. It is how many other plausible objects could satisfy it equally well. If hundreds could, the match remains weak. If only one known object carries the quoted inscription, dimensions, repair and collection label, the match may be highly persuasive even without an image.

Search for contradictions, not only supporting evidence

Confirmation bias is a particular danger when the proposed history is prestigious or valuable. Good matching work actively looks for evidence that would make the identification less likely or impossible. A contradiction should not be hidden inside a persuasive narrative; it should be given the weight its significance deserves.

Silence is not automatically contradiction. A catalogue may omit an inscription, simplify a title or round dimensions. But omission becomes more important when a highly distinctive feature would normally have been recorded, or when the source claims to be a detailed condition or inventory record.

Account for altered, restored and reconstructed objects

An object can be the same historical object without remaining materially unchanged. Cleaning, repainting, rebinding, reframing, trimming, relining, replacement fittings, reconstructed packaging, divided sets and married components can all change the comparison. The task is to describe which kind of continuity survives.

Identity continuity asks whether the historic object remains recognisably the same object. Material continuity asks how much original material survives. Component continuity asks which parts remain from the documented state. Presentation continuity asks whether frame, mount, case or packaging remained associated. These are related but not interchangeable judgements.

A disciplined record-to-object matching workflow

Common ways a near match becomes a false pedigree

Repetition is not corroboration

Use confidence wording that exposes the limits of the match

The conclusion should be proportional to both the strength of the evidence and the job the match is being asked to perform. A private research lead can remain possible. A sales description, insurance schedule, catalogue entry or high-value claim requires clearer support and more explicit attribution.

Controlled wording is not timid. It preserves the difference between fact, evidence and inference, and it prevents a later reader from mistaking the collector's proposed match for an institutional determination.

Preserve enough evidence for the match to be re-tested

A robust collector record should not bury the entire conclusion in one free-text provenance field. Source details, match scope, identifying characteristics, contradictions, confidence, attribution and attachments should remain distinguishable even when they are presented together on screen.

For important online records, keep the institution, object number, stable identifier, access date, relevant wording and an archived copy or screenshot where permitted. Databases are revised, merged, renumbered and sometimes removed. A bare hyperlink is not durable evidence.

What a successful match does - and does not - establish

When specialist or institutional assistance is warranted

Most routine matches can be documented by a careful collector, particularly where the conclusion remains modest. Escalation becomes appropriate when the claim materially affects value, ownership, authenticity, ethics or the interpretation of a complex object.