Attribution Documentation

Attribution documentation records the reasoned assignment of an object - or a particular component, mark, version or association - to a maker, manufacturer, artist, workshop, publisher, factory, place, period, printing, event or person. A defensible record does more than store the attractive name attached to the object. It preserves the exact claim, its scope, the evidence supporting it, the degree of confidence, the alternatives considered and the limitations that remain.

Attribution should therefore be treated as a revisable conclusion rather than a permanent label. New comparators, archive records, technical findings or specialist opinions may strengthen, narrow or overturn the current view. The documentation must allow that change without erasing how the collection arrived at its earlier position.

Collector scenario

The old label that quietly became authorship

A collector acquires an object bearing an old paper label reading "Made by A. Smith, 1890." The first database entry says "Smith attribution from old collection label." Years later an insurance schedule shortens this to "A. Smith, 1890," and a sale draft describes the object as "by A. Smith." No new evidence has appeared; only the wording has strengthened.

A careful record separates the layers: the observable fact is that the label bears those words; the historical source claim is that an earlier owner attributed the object to Smith; the present conclusion may be that the object is period-consistent but the named attribution remains unverified. The label is evidence of a belief. It is not, by itself, proof of authorship.

Attribution is not the same as authenticity

Authenticity asks whether the object is genuinely what it is represented to be. Attribution asks who or what should be credited with its making, design, issue, production, date, place, version or association. The concepts overlap, but one does not automatically settle the other.

A genuine nineteenth-century chair may be attributed to the wrong cabinetmaker. An authentic vintage toy may be assigned to the wrong factory. A genuine early game book may be misdescribed as a first printing. Conversely, the correct manufacturer may be identified even though the signature, label, accessory or claimed celebrity association is later or unsupported.

Authenticity question

Is the object genuinely what the claim says it is?

Original rather than reproduction; period rather than modern; genuinely signed; genuinely issued in the stated production; genuinely retaining the claimed components.

Attribution question

To which maker, source, production or association does it belong?

Which artist, manufacturer, printer, workshop, factory, edition, date range, region, owner, performer, event or issuing body is supported - and with what confidence?

Avoid collapsing object authenticity, maker attribution, date, edition, association, component originality and provenance into a single phrase such as "fully authentic." Each is a different proposition and may rest on different evidence.

Define exactly what is being attributed

Attribution is broader than naming an artist. A single collectible may carry several nested attributions at once, and evidence for one should not be allowed to spill into another. A printed collectible, for example, may have a publisher, printer, designer, cover artist, edition, printing, market, date and former owner. Those are separate fields because they are separate claims.

Maker

Who physically produced it?

This may be an artist, printer, binder, potter, mint, factory, workshop, prop fabricator, contract manufacturer or other producing body.

Do not assume the brand named on the object was the physical manufacturer.

Design

Who created the design?

Design attribution may concern an illustrator, sculptor, industrial designer, game designer, pattern designer or graphic artist even where another party manufactured the object.

Issue

Who authorised or released it?

Publisher, record label, film studio, licensing company, sports club, convention organiser or promotional campaign may be the issuing source without being the maker.

Production

Where in the production system does it belong?

An attribution may identify a factory, pressing plant, print house, foundry, animation studio, production line or prototype phase.

Version

Which edition, printing, issue or variant is it?

First edition, third printing, regional issue, export version, factory variation, test pressing and promotional issue are separate propositions that require their own evidence.

Association

Who owned, used, wore or presented it?

Screen-used, stage-worn, player-used, author-owned, presentation copy and exhibition-displayed claims require object-specific linkage. Correct manufacture does not prove personal use.

Composite objects require component-level scope

The wording should apply only to the part or proposition actually examined. A framed artwork may have an attributed painting and a later frame. A toy may have an original body and replacement weapon. A signed book may contain a genuine signature and an inscription in another hand. A watch may combine an authentic movement, replacement dial and later bracelet.

Prefer: "The painted artwork is attributed to John Smith; the frame is later and has not been attributed." Avoid: "Attributed to John Smith" where the reader cannot tell whether the statement covers the whole assembled object.

Treat attribution as a claim, not a name field

A weak record says only "Manufacturer: Palitoy." A stronger record explains the proposition: "Attributed to Palitoy, Coalville production, approximately 1978-79, based on packaging typography, mould details, country-of-origin marking and comparison with documented examples. No factory record or original purchase documentation has been located."

The second version preserves the source, scope, date, evidence, limitations and uncertainty. This matters because a name entered without its basis can migrate into valuations, insurance schedules, auction descriptions and databases as though it were established fact.

A practical attribution statement formula

[Qualifier] [maker or source], [place or production], [date or range], based on [principal evidence]. [Material limitation or unresolved issue].

"Attributed to Manufacturer X, Japanese production, circa 1977, based on mould markings, packaging construction and comparison with documented retail examples. The exact factory has not been established."

"Published by Company Y and identified as the second printing from the printer's key, advertisement sequence and corrected text on page 14."

"Traditionally attributed as stage-worn by Performer Z. The garment is period-correct, but no photograph, wardrobe record or continuous provenance connects this example to the performer."

Use a confidence vocabulary that has defined meanings

Terms such as "possibly," "probably" and "attributed to" are useful only when the collection uses them consistently. A qualifier should communicate the current strength of the conclusion, not the collector's enthusiasm. Numerical percentages often create false precision unless the field has a reproducible probability method.

Direct

By / manufactured by / published by

Use a direct formulation only where the evidence is strong enough for the principal record to state the attribution without qualification.

Strongly favoured

Probably by / highly probable

The evidence strongly favours one attribution, but direct confirmation or a decisive record remains absent.

Plausible

Attributed to / possibly by

The proposed source is credible enough to record, but the evidence does not discriminate adequately between it and reasonable alternatives.

Production relationship

Workshop of / studio of / circle of

Use only where the field genuinely recognises these production relationships. The terms should not be imported mechanically into categories with different manufacturing structures.

Influence or derivation

After / follower of / school of / style of

These terms describe copying, influence, tradition or resemblance. They do not attribute personal execution to the named maker.

Record status

Formerly attributed / unverified / unresolved / rejected

Use status language to preserve an attribution history without presenting an obsolete or unsupported claim as current fact.

Suggested internal confidence statuses

Status	Working meaning
Confirmed	Direct or convergent evidence strongly supports the attribution and no material contradiction remains.
Highly probable	Evidence strongly favours the attribution, but direct confirmation is absent.
Probable	The attribution is more likely than the competing explanations.
Possible	Evidence is consistent with the attribution but is not sufficiently discriminating.
Unresolved	Two or more plausible attributions remain.
Unsupported	A claim exists, but no adequate supporting evidence has been recorded.
Contradicted	Material evidence conflicts with the claim.
Rejected	The attribution has been examined and is no longer accepted.
Not assessed	No attribution assessment has yet been undertaken.

Collector risk

Prestige is not a method. A certificate or important collection may merely repeat the same untested claim supplied by the seller.

Make the reasoning chain visible

Step 1

Observation

The copyright line contains a particular address.

Step 2

Reference

The publisher used that address from 1982 to 1984.

Step 3

Interpretation

The object was probably printed within or after that period.

Step 4

Consequence

The seller's claimed 1979 printing is unlikely.

Distinguish direct, indirect and contextual evidence

Direct

Explicitly links this object to the source

A factory record listing the serial number, a photograph showing the unique object in use, an original invoice describing it or an archive entry matching its identifying features.

Indirect

Supports through discriminating inference

Construction, materials, handwriting, dimensions or printing details that match secure examples and make one attribution more likely than alternatives.

Contextual

Shows historical possibility

The maker worked in the relevant place, the company held the licence or the person attended the event. Context establishes plausibility, not individual identity.

Separate observed fact, source statement and present conclusion

Attribution errors often begin when another person's wording is copied into the collection record without a prefix. The catalogue should distinguish what can be seen, what an earlier source claimed and what the collection currently concludes.

Observed fact

A paper label attached to the underside bears the handwritten wording 'Made by A. Smith, 1890.'

Source statement

The label has been attributed to former owner B. Jones, but this has not been independently confirmed.

Present conclusion

The object is currently attributed to an unidentified regional workshop, circa 1880-1910. The A. Smith attribution remains unverified.

Useful prefixes include "inscribed," "labelled," "described by the seller as," "recorded in the 1975 inventory as," "traditionally attributed to," "identified by X as" and "independently assessed as." These small phrases stop a source's claim from becoming the catalogue's endorsement.

Test competing hypotheses, not only the desired answer

Good attribution records are designed to expose confirmation bias. For a collectible that might be an original release, a licensed reissue, a later reproduction or a composite assembled from different generations, record how each observation affects each possibility. The purpose is not to create a decorative matrix; it is to show that plausible alternatives were actively tested.

Compact hypothesis test

Observation	1978 original	1980 reissue	Composite
Body mould	Supports	Supports	Supports
Country mark	Supports	Conflicts	Supports
Accessory colour	Conflicts	Supports	Supports
Packaging print	Supports	Conflicts	Neutral
Age consistency	Supports	Supports	Mixed

Watch for attribution problems that change by category

People and personal association

Personally designed, personally made, supervised, signed, owned, used, worn, presented and merely associated are materially different claims. A commercially manufactured guitar may be correctly attributed to Company Y while its claimed ownership and use by Musician X remain unresolved.

Corporate and licensed production

A company name may identify the brand owner, licence holder, distributor, importer, retailer, tooling owner or successor business rather than the manufacturer. Use historically correct entity names and date ranges, especially where trademarks, tooling or contracts moved between companies.

Editions, printings and issues

First edition and first printing are not synonyms. Printer's keys, colophons, addresses, advertisements, corrected errors, matrix numbers, labels, paper and regional legal text may be needed. A useful record may say 'first edition, third printing - not a first printing.'

Signatures and inscriptions

Document the signature separately from the object, dedication, date and annotations. A genuine object may carry a later forged addition; a genuine signature may appear on an unrelated object. Record exactly what was examined and whether other writing appears to be in another hand.

Style-based attribution

Style can support attribution through composition, typography, modelling, colour, ornament or design habits, but it is vulnerable to imitation, workshop repetition, common period conventions and restoration. Name the observed features instead of writing only 'the style is right.'

Marks and codes

Transcribe marks exactly, including punctuation, letter case, orientation, location, size, production method and uncertain characters. Identify whether a date range applies to the company, the trademark, the mark design, the site or the object type.

Domain boundary

Attribution, provenance, description and valuation must remain separate

Attribution identifies likely origin, authorship, production or association. Provenance records ownership and custody. Description records observable features. Authentication tests whether the claimed identity is genuine. Grading evaluates condition or quality. Valuation may depend on attribution, but price is not evidence that the attribution is correct.

These domains inform one another but should not become circular. "It came from an important collection, so it must be by Maker X" followed by "it is by Maker X, which proves the collection was important" is not independent reasoning.

Record restoration before relying on attribution features

Restoration can obscure, remove or manufacture the very features used for attribution. Overpainting may hide original brushwork; replacement labels may create false maker associations; refinishing can remove stamps; reconstructed signatures can appear stronger than the surviving original; replacement components can carry incompatible marks; rebacking or relining can conceal paper evidence and inscriptions.

Before relying on a feature, document:

what is original, restored, reconstructed or replaced;
which attribution features remain visible and accessible;
whether the attribution predates or postdates treatment;
whether treatment records were reviewed;
whether cleaning, coating or repair changed how the evidence appears.

Certificates and expert opinions are evidence, not substitutes for the record

A useful certificate uniquely identifies the object, states who examined it, when and by what method, gives the attribution wording and confidence, summarises the evidence, records limitations and can be verified. A certificate that contains only a famous name, stock photograph and unexplained conclusion may be commercially impressive but intellectually weak.

Strong certificate features

issuer identity, qualifications, date and unique certificate number;
precise object description, dimensions, marks and object-specific photographs;
scope and conditions of examination;
attribution wording, confidence, evidence summary and limitations;
signature or verifiable digital authentication and reassessment terms.

Warning signs

no date, issuer identity or method;
generic photographs that do not uniquely identify the object;
certainty language unsupported by reasoning;
opinion issued only from seller-supplied images without saying so;
a certificate that could be transferred between similar objects.

Remote attribution must disclose what could not be examined

Photo-based review can be valuable for triage, comparison and deciding whether specialist examination is justified. It may not reveal weight, texture, internal construction, fluorescence, concealed marks, layer structure, smell, sound, material composition or whether every photograph shows the same object. Preserve the supplied files, date received and identity of the source.

Suggested wording

"Preliminary attribution based on supplied digital photographs. No in-person examination, measurement or material analysis has been undertaken."

Preserve disagreement and attribution history

When recognised experts disagree, the record should not hide the less commercially attractive opinion. Record each conclusion, who made it, when, what evidence was cited, whether the object was examined in person, the methodological difference and what new evidence might resolve the dispute. "Disputed" is often more accurate than selecting whichever name produces the highest value.

Attribution history example

Date	Attribution	Status	Basis
1974	Maker A	Previous	Unsourced auction catalogue statement
1998	Workshop of Maker A	Superseded	Stylistic comparison by Curator B
2024	Unknown maker, Region X	Current	Material study and revised archive evidence

Prevent attribution laundering

Attribution laundering occurs when a weak claim gains authority through repetition. A seller writes "possibly associated with X"; a later listing shortens it to "associated with X"; an auction catalogue says "from the X production"; a database enters "used by X"; the database is then cited as confirmation. No new evidence has entered the chain, but the claim appears more certain at every stage.

Documentation that interrupts the laundering sequence

preserve the earliest located wording and its original source;
identify later sources that depend on that first statement;
record every change in qualifier or scope;
state whether any new object-specific evidence was added;
treat dependent repetition as one attribution tradition, not multiple confirmations.

The attribution record a serious collector should keep

The aim is not to create an impressive mass of paperwork. It is to preserve enough observation, source information and reasoning for another competent person to evaluate the conclusion independently. Routine objects may need a concise record; high-value, disputed or historically important objects need a fuller attribution file.

Identify the object and the claim

Collection or accession number, title, object type and current location.
Photographs, dimensions, serial numbers, marks and component inventory.
The exact target: maker, designer, factory, publisher, place, date, edition, printing, event association, owner or user.
The exact component or proposition covered by the attribution.

State the present conclusion

A concise attribution statement using a defined qualifier.
A confidence status that has an agreed meaning within the collection.
The principal reasons supporting the conclusion.
Any material counter-evidence, contradiction or unresolved issue.

Make the reasoning auditable

Full source references and access details where appropriate.
Comparator records and why those examples are considered secure.
Observation, reference, interpretation and attribution consequence.
Competing hypotheses considered and why they were accepted or rejected.

Record examination and limitations

Examiner, date, method, equipment and whether review was in person or remote.
Components examined and any areas concealed, framed, restored or inaccessible.
Tests not undertaken and archives or comparators that were unavailable.
Whether the attribution depends on a restored, replaced or reconstructed feature.

Preserve responsibility and change

Who made or approved the attribution.
Date assigned, dates reviewed and reasons for revision.
Previous, superseded, disputed or rejected attributions.
A review trigger, such as new archive records, a technical report or access to the object in person.

Minimum private-collector record

At minimum, record the attribution exactly as claimed; who made the claim and when; what part of the object it concerns; the confidence level; supporting and contrary evidence; photographs of identifying features; source references; and the date and author of the present conclusion.

Current attribution: Probably a 1983 UK-market release manufactured under licence for Company X.

Basis: UK legal text, package dimensions, logo form and accessory configuration match three documented examples.

Claim source: Sold in 2019 as a 1981 first release; no supporting documentation supplied.

Contrary evidence: Package address and revised safety wording were introduced after 1982.

Review trigger: Reassess if dated trade catalogues or factory records become available.

When specialist escalation is proportionate

A collector can reasonably photograph marks, transcribe labels, trace claim sources, compare secure references, separate components and write a qualified conclusion. Seek specialist help when the proposed attribution materially changes value, legal description, insurance, export treatment, catalogue inclusion or historical significance; when technical analysis is needed; when restoration conceals decisive features; or when recognised experts disagree.

Collector-level action

record the exact claim and its source;
photograph and transcribe marks without normalising them;
separate observations from interpretation;
identify secure comparators and plausible alternatives;
use cautious wording and preserve uncertainty.

Specialist threshold

the attribution has major financial or legal consequences;
material, ink, pigment, alloy, paper or imaging analysis is required;
the object must be dismantled or examined beneath restoration;
the claimed association is celebrity, screen, stage, sports or event use;
the existing record is disputed, contradictory or likely to enter the market.

Common failures that weaken attribution records

Unsupported certainty

A named attribution is entered with no evidential basis or reasoning.

Resemblance alone

The object looks similar, but no discriminating features or competing makers are tested.

Seller wording becomes fact

A sales description is copied directly into the collection record without independent assessment.

A mark is treated as conclusive

Copied, transferred, altered, reused or component-specific marks are not considered.

Brand and manufacturer are conflated

The company named on the product is assumed to have physically produced it.

Maker and user are conflated

Production for a film, tour or team is treated as proof of use by the named individual.

Restored areas are ignored

The attribution depends on a feature that is partly or wholly reconstructed.

Consensus is overvalued

Dependent repetitions are mistaken for independent corroboration.

Old opinions are deleted

Superseded attributions disappear, leaving earlier labels and values unexplained.

Value influences identity

The most valuable attribution is preferred to the best-supported conclusion.

Collector review questions

What exactly am I attributing: manufacture, design, signature, ownership, use, date, place or version?
Is the evidence object-specific, or does it identify only a general type?
Which observations could another person verify directly?
What is the earliest source for the claim?
Are later sources independent, or are they repeating one tradition?
Are the comparators securely attributed?
What alternative explanations fit the evidence?
Is any decisive feature restored, replaced, obscured or inaccessible?
Does the wording communicate the real uncertainty?
Could the conclusion be reassessed from the record alone?
Would a buyer understand what is established and what remains claimed?

Central rule

Never allow an attribution to become detached from its qualifier, source, scope and evidential basis.

"Attributed to," "probably by," "workshop of," "signed by," "owned by" and "used by" describe materially different propositions. A defensible record does not merely answer "Who made this?" It records what is being attributed, to whom or what, by whose judgement, on what evidence, with what confidence, against which alternatives, subject to which limitations and with what history of revision.

Continue learning

Authenticity Documentation

Review how to document what has actually been authenticated and where limits remain.

→

Back to Authentication Documentation

Return to the Authentication Documentation sub-domain and its full topic list.

→

Probability & Confidence Documentation

Continue to recording confidence levels, uncertainty and what would change the conclusion.

→