MUSEUM OBJECT IDENTITY / ATTRIBUTION / PROVENANCE DOSSIER v0.1

Status: RESEARCH / PRE-CAPTURE ONLY
Utility family: PERSISTENT OBJECT IDENTITY + VERSIONED DESCRIPTION
Fixture: National Gallery of Art accession 1942.9.70, Portrait of Rembrandt, dated 1650
Research date: 2026-08-09
Authority: NONE — research input only
Explicit non-claims: No CaptureReceipt, SourceEdition, SRS receipt, attribution adjudication, ownership determination, standing, or canonical Counterpedia identity is created here.

A. WHY THIS SUBJECT

The user question is calm:

“What is this painting?”

A conventional encyclopedia answers with a title and artist.

But a museum object can outlive many changing descriptions:

• who the sitter is;

• who painted it;

• what the title is;

• when it was made;

• what collection it belonged to;

• where it was physically located;

• how conservators understood its materials;

• what a catalogue considered authentic;

• whether a signature was accepted;

• and which image reproduction or catalogue entry a reader saw.

This fixture demonstrates a new Counterpedia primitive:

A persistent physical object is not identical to any one description of it.

The object can remain the same while its metadata changes around it.

B. FIXED OBJECT ANCHOR

ART-OBJ1 — Physical painting

Current institutional title: Portrait of Rembrandt
Current institutional attribution: Dutch 17th Century
Date: 1650
Medium: oil on canvas
Accession number: 1942.9.70
Credit line: Widener Collection
Owning institution: National Gallery of Art, Washington

The accession number is the most useful institutional anchor in this dossier.

It lets Counterpedia distinguish:

physical object
    ≠
current title
    ≠
current attribution
    ≠
historical title
    ≠
historical attribution
    ≠
current physical location
    ≠
digital image

The accession number itself is an institutional identifier, not a universal ontology ID.

C. CURRENT INSTITUTIONAL STATE

The National Gallery of Art currently describes the painting as:

title: Portrait of Rembrandt
artist/attribution: Dutch 17th Century
date: 1650
medium: oil on canvas
accession: 1942.9.70

The NGA states that the painting was once celebrated as a Rembrandt self-portrait, but current technical and stylistic analysis indicates that it was painted by an unknown artist in Rembrandt's workshop rather than by Rembrandt himself.

As observed in August 2026, the NGA object page also says the painting is on loan to the Denver Art Museum as part of Across the Nation.

This creates another useful distinction:

institutional ownership / accession
    ≠
current physical custody / display location

A loan does not create new ownership.

D. DESCRIPTION HISTORY

The same object has carried materially different descriptions.

D1 — Early nineteenth century: Admiral Tromp

In the Érard collection, the sitter was identified as the Dutch admiral Maerten Harpertsz Tromp.

The 1832 estate-sale provenance entry identifies the work as:

Portrait de Martin-Kappertz-Tromp

William Williams Hope's 1849 sale likewise described it as a portrait of Admiral Van Tromp.

The current NGA scholarship states that John Smith was the first to correctly identify the sitter as Rembrandt.

Therefore:

sitter_identity_claim
  Tromp
      ↓ superseded
  Rembrandt

The object did not change.

The description did.

D2 — Nineteenth and early twentieth centuries: Rembrandt self-portrait

Once the sitter was recognized as Rembrandt, the painting was widely treated as a self-portrait painted by Rembrandt.

NGA's bibliography records repeated catalogue descriptions such as:

• Self-Portrait by Rembrandt;

• Portrait of Himself by Rembrandt;

• inclusion in Rembrandt catalogues and exhibitions.

This was not merely a casual nickname. It was a durable scholarly/institutional attribution state.

D3 — 1969: Gerson rejects Rembrandt authorship

In Horst Gerson's 1969 revision of Bredius's Rembrandt catalogue, the painting was rejected as an authentic Rembrandt and characterized as a later imitation.

NGA's scholarly entry describes the reaction as a major reassessment because the work had long been accepted and admired as an important self-portrait.

This is a critical Counterpedia state:

historical scholarly attribution
  Rembrandt
      ↓
major catalogue reassessment
      ↓
not Rembrandt

But Gerson's specific alternative description — later imitation — did not itself remain the final institutional state.

D4 — Technical study: not a later imitation

NGA technical analysis found that:

• paint mixtures;

• pigments;

• canvas;

• double ground;

• buildup;

• execution changes visible in X-radiographs

are consistent with Rembrandt's workshop practices.

The NGA scholarly entry explicitly says this technical evidence shows Gerson was wrong to call it an eighteenth- or nineteenth-century imitation.

So:

Gerson:
  later imitation
      ↓ contradicted by technical evidence
NGA:
  seventeenth-century workshop context

This is not the same as restoring the painting to Rembrandt.

D5 — Current attribution: unknown artist in Rembrandt's workshop

NGA's current position is more precise:

• stylistic differences are sufficient to remove the painting from Rembrandt's own oeuvre;

• the actual painter cannot be determined;

• the painting was probably executed by a member of Rembrandt's workshop around 1650;

• no known workshop painter can be convincingly associated with it.

This gives the final sequence:

OBJECT 1942.9.70

historical sitter:
  Admiral Tromp
      ↓
  Rembrandt

historical artist:
  Rembrandt
      ↓
  later imitator (Gerson)
      ↓
  unknown 17th-century Rembrandt-workshop artist

current institutional label:
  Dutch 17th Century

The accessioned object survives all of those changes.

E. PRIMARY SOURCE INVENTORY

ART-S01 — Current NGA object page

Issuer: National Gallery of Art
Title: Portrait of Rembrandt
Locator: https://www.nga.gov/artworks/1209-portrait-rembrandt
Source class: first-party collection record
Priority: P0

Required facts:

• current title;

• current institutional attribution Dutch 17th Century;

• 1650;

• oil on canvas;

• Widener Collection;

• accession 1942.9.70;

• current summary of workshop attribution;

• current loan/display state;

• provenance;

• exhibition history;

• bibliography.

Important temporal field The current loan statement is operational/current-state metadata and can change independently of the object's title, ownership, or attribution.

ART-S02 — NGA scholarly catalogue entry

Issuer: National Gallery of Art
Series: Dutch Paintings of the Seventeenth Century
Title: Portrait of Rembrandt, 1650
Author: Arthur K. Wheelock Jr.
Publication date: April 24, 2014
Locator: https://www.nga.gov/research/publications/online-editions/dutch-paintings-seventeenth-century-portrait-rembrandt-1650
Source class: first-party museum scholarly catalogue
Priority: P0

This is the main attribution-history source.

Required facts:

• early Tromp sitter identification;

• later recognition of Rembrandt as sitter;

• historical acceptance as Rembrandt self-portrait;

• Gerson's 1969 disattribution;

• technical evidence rejecting Gerson's later imitation characterization;

• stylistic reasons for removing the work from Rembrandt's own oeuvre;

• current workshop conclusion;

• unresolved painter identity;

• technical summary;

• conservation treatment history.

ART-S03 — NGA Rembrandt attribution methodology essay

Issuer: National Gallery of Art
Title: Issues of Attribution in the Rembrandt Workshop
Author: Arthur K. Wheelock Jr.
Published online: June 1, 2014
Locator: https://www.nga.gov/research/publications/dutch-paintings-seventeenth-century/issues-attribution-rembrandt-workshop
Source class: first-party scholarly methodology / collection interpretation
Priority: P1

This source places the object inside a broader institutional history:

• early Rembrandt catalogues accepted very large corpora;

• later scholarship reduced the oeuvre;

• Gerson's 1969 catalogue rejected this painting along with other famous works;

• attribution depends on provenance, historical descriptions, technical imaging, conservation science, visual comparison, and scholarship.

Counterpedia value:

Attribution is an evidence-backed institutional proposition, not a permanent property magically attached to canvas.

ART-S04 — Provenance section of current NGA collection record

Issuer: National Gallery of Art
Source object: ART-S01 section-level record
Priority: P0

Current NGA provenance chain begins with:

• Chevalier Sébastien Érard;

• Érard estate sale, 1832;

• William Williams Hope by 1836;

• sale in 1849;

• Sir Anthony Nathan de Rothschild;

• Lady Anthony de Rothschild;

• Thomas Agnew & Sons;

• sold 13 May 1908 to Peter A. B. Widener;

• Widener family succession;

• gift to NGA in 1942.

Important boundary:

The NGA provenance record does not claim a complete chain back to 1650. Its scholarly note says provenance is not known before the Érard collection.

Therefore:

provenance_known_from ≈ early 19th century
≠
provenance_complete_since_creation

ART-S05 — Current NGA collection image / digital surrogate

Issuer: National Gallery of Art
Available from: ART-S01 / NGA image API
Source class: institutional digital reproduction
Priority: P0

This source should be represented separately from the painting.

The digital image has:

• bytes;

• dimensions;

• crop;

• color-management/reproduction characteristics;

• publication/capture time;

• image-service identity.

It is evidence of the object but is not the physical painting.

Counterpedia should support:

digital_image
  depicts
physical_object

not:

digital_image
  is
physical_object

ART-S06 — Technical image family

Issuer: National Gallery of Art
Source family: X-radiography / conservation images referenced in ART-S02
Priority: P1

The catalogue states that X-radiographs reveal changes made during execution, including changes to the skullcap/beret.

These are not decorative images.

They are evidence objects supporting claims about the work's manufacture.

ART-S07 — Conservation-treatment record

Issuer: National Gallery of Art
Source: technical summary in ART-S02; deeper conservation files may exist institutionally
Priority: P1

The scholarly entry records treatment in 1992–1993, including removal of discolored varnish and later inpainting/overglaze.

Counterpedia value:

A physical object's visible state can change because of conservation.

Therefore:

appearance_at_time_t1
≠
appearance_at_time_t2

without implying a different painting.

F. PROVENANCE IS A CHAIN, NOT A PARAGRAPH

A traditional object page renders provenance as punctuation-heavy prose.

Counterpedia should model events.

Example:

OwnershipEvent
  object = ART-OBJ1
  holder = Sébastien Érard
  valid_before = 1832

TransferEvent
  kind = estate_sale
  date = 1832-08
  catalogue_lot = 119
  title_at_event = Portrait de Martin-Kappertz-Tromp

OwnershipEvent
  holder = William Williams Hope
  known_by = 1836

TransferEvent
  kind = auction
  date = 1849-06
  title_at_event = Portrait of Admiral Van Tromp

...

TransferEvent
  kind = sale
  date = 1908-05-13
  recipient = Peter A. B. Widener

TransferEvent
  kind = gift
  date = 1942
  recipient = National Gallery of Art

This makes provenance queryable:

Where was it in 1909?

Under what title was it sold in 1832?

Who owned it before Widener?

Where does the known chain begin?

G. PROVENANCE GAPS ARE FIRST-CLASS

The current scholarly record says the painting's provenance is not known before the Érard collection.

Counterpedia should render:

1650 ───────────────────── ? ───────────────────── early 1800s
                           ↑
                  provenance not established

Not:

1650 → Érard

A visually continuous line would manufacture knowledge.

The gap itself is part of the record.

H. ATTRIBUTION AS VERSIONED STANDING

Counterpedia should not overwrite the attribution field.

Use historical states.

Conceptually:

AttributionState A1
  valid_in_catalogue_context = early 19th c.
  sitter = Admiral Tromp
  artist = Rembrandt

AttributionState A2
  valid_in_scholarship_context = later 19th / early 20th
  sitter = Rembrandt
  artist = Rembrandt
  object_role = self_portrait

AttributionState A3
  source = Gerson 1969
  artist = not Rembrandt
  characterization = later imitation

AttributionState A4
  source = NGA technical/scholarly assessment
  artist = unknown workshop artist
  period = 17th century
  sitter = Rembrandt

CurrentInstitutionalProjection
  title = Portrait of Rembrandt
  attribution = Dutch 17th Century

Historical state is preserved even after supersession.

I. ATTRIBUTION IS MULTI-DIMENSIONAL

Do not model:

artist = ?

alone.

This fixture contains separate propositions:

sitter_identity
creator_identity
workshop_context
creation_date
signature_authorship
signature_integrality
object_role

The current NGA scholarship says, in substance:

• sitter is Rembrandt;

• creator is not accepted as Rembrandt;

• creator is unknown;

• workshop context is Rembrandt's;

• date is about 1650;

• signature/date are integral to the paint surface but apparently not written by Rembrandt.

Those are different edges.

J. SIGNED "REMBRANDT" ≠ PAINTED BY REMBRANDT

The painting bears a signature/date reading Rembrandt f. 1650.

Yet the current institutional attribution removes it from Rembrandt's own oeuvre.

This provides a quiet, memorable refusal:

“It literally says Rembrandt on the painting, so Rembrandt painted it.”

Counterpedia should answer:

The signature is one piece of physical evidence. NGA's current scholarly assessment says the signature appears integral to the surface but apparently was not written by Rembrandt, while stylistic evidence excludes the painting from his own oeuvre.

This demonstrates:

inscription text
≠
authorship adjudication

K. TECHNICAL EVIDENCE CAN REJECT ONE THEORY WITHOUT RESTORING AN OLDER ONE

This is one of the best epistemic structures in the fixture.

Gerson's 1969 theory:

not Rembrandt
AND
18th/19th-century imitation

Later technical analysis supports:

not later imitation

But that does not entail:

therefore Rembrandt

Instead the current state is:

17th-century Rembrandt workshop
+
unknown painter
+
not Rembrandt's own oeuvre

The inference structure is:

theory A rejected
≠
previous theory B restored

This is the museum-object equivalent of the curb rule:

removing one explanation does not create another automatically.

L. CONSERVATION AS OBJECT HISTORY

The physical object itself has a history.

NGA records:

• canvas support and grounds;

• later lining;

• paint layers;

• abrasions/losses;

• X-radiographic pentimenti/changes;

• 1992–1993 conservation treatment;

• removal of discolored varnish and later overpaint.

Counterpedia should support:

PhysicalState
  object
  observed_at
  imaging_method
  condition
  treatment_history

This makes possible:

What did conservators discover in 1993 that changed attribution?

rather than only:

Who painted it?

M. EXHIBITION ≠ OWNERSHIP

The object has appeared in exhibitions including:

• Royal Academy of Arts, London, 1899;

• Metropolitan Museum of Art, New York, 1909;

• National Gallery of Art Rembrandt exhibition, 1969.

As of the August 2026 NGA page, it is on loan to the Denver Art Museum.

These are events:

ExhibitionEvent
LoanEvent
LocationState

not ownership transfers.

A useful object page needs both:

owned_by = NGA
located_at = Denver Art Museum (current loan state)

with dates.

N. DIGITAL OBJECT ≠ PHYSICAL OBJECT

This fixture should explicitly show three layers:

PHYSICAL OBJECT
  accession 1942.9.70

DOCUMENTARY RECORD
  NGA object page
  scholarly catalogue
  old sale catalogue
  exhibition catalogue

DIGITAL SURROGATE
  JPEG/TIFF/image-service rendition
  X-ray image
  web thumbnail

A reverse-source page for a JPEG should be able to say:

this image depicts ART-OBJ1

without giving the JPEG the painting's ownership/provenance.

This matters enormously once Counterpedia becomes machine-readable.

O. REVERSE-WIKIPEDIA VALUE

Open the 2014 NGA scholarly catalogue.

Reverse:

technical analysis
  pigments + grounds + canvas + X-ray
        ↓
rejects "later imitation"
        ↓
supports 17th-c workshop context
        ↓
stylistic analysis
        ↓
excludes Rembrandt authorship
        ↓
current attribution state

Open the 1832 provenance record:

sale catalogue
  title = Portrait de Martin-Kappertz-Tromp
        ↓
historical sitter claim
        ↓
later sitter identification supersedes it

The source page can answer:

Which current and historical claims depend on this old catalogue?

That is Reverse Wikipedia applied to cultural heritage.

P. FIVE QUIET-UTILITY CLAIM ROLES

These are research vocabulary candidates, not frozen global enums.

ART-Q1 — PERSISTENT OBJECT IDENTITY

NGA accession 1942.9.70 identifies the same institutional physical object across changing catalogue descriptions.

ART-Q2 — HISTORICAL DESCRIPTION

The object was sold in the 1832 Érard estate sale under a title identifying the sitter as Admiral Tromp.

This is true about the historical record, even though the sitter identification is no longer accepted.

ART-Q3 — CURRENT ATTRIBUTION

NGA currently attributes the painting to Dutch 17th Century and describes it as the work of an unidentified artist from Rembrandt's workshop.

ART-Q4 — SUPERSEDED ATTRIBUTION

The painting was historically accepted as a Rembrandt self-portrait.

This remains a valid historical claim.

It must not be deleted merely because the attribution changed.

ART-Q5 — PROVENANCE GAP

NGA does not establish the painting's provenance before the Érard collection.

The absence is a positive statement about the limits of the current record.

Q. WHY-NOT SPECIMENS

Why not 1

“The museum changed the artist, so it must be a different painting.”

Refuse.

The accessioned physical object is persistent; attribution is a changing proposition about its authorship.

Why not 2

“It says ‘Rembrandt f. 1650,’ so Rembrandt painted it.”

Refuse.

The inscription is physical evidence, but current NGA scholarship does not accept Rembrandt as the painter.

Why not 3

“Gerson proved it was a nineteenth-century fake.”

Refuse.

Gerson proposed a later-imitation theory in 1969. NGA's technical analysis later rejected that dating/context while still excluding Rembrandt authorship.

Why not 4

“The provenance says who owned it since 1650.”

Refuse.

NGA explicitly notes that provenance is not known before the Érard collection.

Why not 5

“It's at the Denver Art Museum, so Denver owns it now.”

Refuse.

Current location/loan state and institutional ownership are different relations.

R. 30-SECOND DEMO

Open a beautiful image of the painting.

Counterpedia says:

OBJECT
NGA 1942.9.70
Portrait of Rembrandt
1650

Then a timeline:

1832

SITTER
Admiral Tromp

ARTIST
Rembrandt

Slide forward:

early 20th century

TITLE / ROLE
Self-Portrait

ARTIST
Rembrandt

Slide:

1969

ARTIST
not Rembrandt

Gerson:
later imitation

Slide:

CURRENT NGA

SITTER
Rembrandt

ARTIST
Dutch 17th Century
unknown Rembrandt-workshop painter

OBJECT
still NGA 1942.9.70

Then click:

What changed?

Counterpedia:

not the painting

changed:
  sitter identification
  authorship attribution
  scholarly explanation
  conservation evidence
  catalogue state

The physical object stayed put in the identity graph.

That's the reveal.

S. SECOND DEMO: WHERE IS IT?

Ask:

Where is the painting?

Counterpedia can answer two different questions:

OWNED / ACCESSIONED BY
National Gallery of Art

CURRENT PHYSICAL LOCATION
Denver Art Museum
loan state observed on NGA page, Aug 2026

A conventional database often has one museum field.

Counterpedia needs relations.

T. THIRD DEMO: WHAT DID A PERSON IN 1931 "KNOW"?

Ask:

If I opened a respected catalogue in 1931, what would it tell me this object was?

Counterpedia can replay the historical documentary state:

sitter: Rembrandt
artist: Rembrandt
role: self-portrait
standing-in-that-source-context: accepted

Then:

What does the NGA say now?

Current state appears alongside it.

This is temporal replay without calling earlier scholarship dishonest.

U. PRODUCT PRIMITIVES THIS FIXTURE REQUIRES

1. Persistent physical-object identity

Independent of:

• title;

• artist attribution;

• location;

• digital representation.

2. Description assertions with validity/source context

DescriptionAssertion
  object
  predicate
  value
  source
  source_edition
  authored_at
  standing

3. Attribution lineage

supersedes
contradicts
qualifies
restores

but never silent overwrite.

4. Provenance events

Ownership, custody, sale, gift, inheritance, loan, exhibition must be separate.

5. Provenance gaps

Unknown intervals are represented explicitly.

6. Physical-state / conservation events

Treatment and technical examination alter knowledge about the same persistent object.

7. Representation edges

digital_image → depicts → physical_object
xray_image → examines → physical_object
catalogue_record → describes → physical_object

V. CAPTURE PRIORITY — ART CAP1

P0

1. ART-S01 current NGA object page

2. ART-S02 exact 2014 NGA scholarly catalogue entry

3. ART-S04 provenance section / structured current record

4. ART-S05 current NGA high-resolution digital image with exact image-service metadata

5. current NGA object metadata/API representation if separately addressable

P1

1. ART-S03 NGA Rembrandt attribution-methodology essay

2. ART-S06 X-radiographic/technical image artifacts

3. ART-S07 conservation-treatment evidence accessible through NGA

4. exact 1969 exhibition/catalogue record if institutionally available

5. exact historical Widener/NGA catalogue editions

P2

1. recover exact 1832 Érard sale catalogue bytes

2. exact 1849 Hope sale catalogue bytes

3. 1969 Gerson catalogue page, subject to rights/access

4. exhibition records for 1899 / 1909 / 1969

5. current Denver loan/exhibition record from receiving institution

W. CAPTURE / INGEST DISCIPLINE

museum object record
scholarly catalogue
historic catalogue
image artifacts
technical images
        ↓
CAP1-A exact custody
        ↓
CAP1-B dagr-ingest
        ↓
SRC1 documentary identity
        ↓
object-identity resolution
        ↓
description / attribution / provenance assertions
        ↓
ADJ1
        ↓
timeline / provenance / current-state projection

Capture never decides authorship.

The image pipeline never turns a JPEG into the physical painting.

Attribution is downstream governed interpretation.

X. REGRESSION TESTS LEARNED FROM THIS FIXTURE

ART-R1 — Description change does not change physical identity

Same accessioned object survives title/artist changes.

ART-R2 — Historical false/superseded description remains historically true as a documentary claim

1832 catalogue called sitter Tromp remains valid even though sitter = Tromp is not current standing.

ART-R3 — Rejecting a theory does not restore the previous theory

Technical evidence rejects Gerson's later-imitation theory without restoring Rembrandt authorship.

ART-R4 — Signature is evidence, not authorship oracle

Inscription and creator attribution are separate predicates.

ART-R5 — Provenance gaps remain gaps

Do not draw a continuous chain through unknown years.

ART-R6 — Ownership, custody, exhibition and current location are separate relations

A loan does not transfer ownership.

ART-R7 — Digital surrogate is not object

Bytes can depict the object without inheriting its provenance.

ART-R8 — Conservation can change knowledge and appearance without changing identity

Physical treatment belongs to object history.

Y. PROVISIONAL THESIS

This Quiet Utility entry establishes:

Counterpedia can preserve the identity of a physical thing while allowing every description around that thing to evolve independently.

Its strongest public sentence is:

The museum can change who it says painted a painting without changing which painting it is.

Its deeper architecture sentence is:

Object identity, descriptive metadata, attribution, provenance, custody, exhibition, conservation, and digital representation are different relations that happen to converge on one physical object.

And its quiet-corpus invariant is:

A durable knowledge system should preserve the thing while letting our descriptions of the thing change.