Skip to main content
Thumbu Sammy, identified through a captioned duplicate print
Zella, Jess, and Jesse Power, identified later by a family member

Nostalgia - Living Archive · The Museum

Investigation 02 · Three corrected records

The archive got it wrong. Then the evidence changed.

A catalog is not a verdict. Compare what each record said, what it missed, what arrived later, and what a careful revision still cannot claim.

Cases
Ellis Island and the FSA archive
Method
Old claim → evidence → revision → limit
Time
About eight minutes · no account

Before opening a case

Wrong can mean three different things.

Incomplete
Another surviving object carries information this record lost.
Hidden
The source contains evidence that a copy or interface failed to carry.
Revised
A later contributor supplies evidence the original maker never recorded.

Case files · Compare in any order

Do not erase the old record. Show what changed it.

Each case keeps four layers separate: the first record, the problem, the new evidence, and the bounded revision.

01A duplicate carried moreThe photograph stayed the same. The record gained a person.When is a filing label evidence, but not an identity?
Thumbu Sammy, aged seventeen, photographed in profile at Ellis Island in April 1911
New York Public Library · Image ID 418056
First record

One public copy circulated under the historical filing title “Hindoo boy”. It described how the portrait had been catalogued, but supplied no name.

What was wrong

A descriptive or outdated archive title can become mistaken for the person’s identity when the record and the sitter are treated as the same thing.

Evidence that changed it

Another print carried a fuller caption: Thumbu Sammy, aged 17, ex S.S. Adriatic, April 14, 1911. A documented visual comparison connected the prints as the same negative.

Responsible revision

The combined record supports the name Thumbu Sammy, an age of seventeen in the duplicate caption, the ship, and the date. The earlier title remains visible as catalog history, not as his name.

Still outside the evidence

The evidence does not identify the caption writer, establish which words Sammy used for himself, or recover his life before and after the portrait.

02Our copy dropped a fieldWe called her unnamed. The source record already knew her.Can an interface mistake look like an archival absence?
Florence Thompson with two of her children at the pea pickers camp in Nipomo, California, March 1936
Library of Congress · item 2017762903
First record

Our first plaque said this frame did not name its mother because the title described only a migrant worker’s family and seven hungry children.

What was wrong

Our importer had captured title, date, contributor, and rights, but not the Library of Congress summary field. We mistook a gap in our copy for a gap in the archive.

Evidence that changed it

A direct check of the live Library record found the summary naming Florence Thompson. The source had the answer; our data path had hidden it.

Responsible revision

The plaque now names Florence Thompson and says that the names arrived later, rather than claiming that this particular record still left her anonymous.

Still outside the evidence

This correction does not make every person in the frame named, nor does a later summary tell us exactly when or by whom each detail entered the record.

03A family member wrote inA relative turned “drought refugees” back into a family.Who is allowed to improve a public record?
Zella McCann Power, Jess Power, and their baby Jesse camping by a California roadside in August 1936
Library of Congress · item 2017763095
First record

The photograph’s title describes drought refugees from Oklahoma camping by the roadside and hoping to work in the cotton fields.

What was wrong

The condition was recorded more durably than the people. Read alone, the title turns a specific family into a representative scene.

Evidence that changed it

The Library record now adds Zella (McCann) Power, her husband Jess, and their baby Jesse, and credits the identification to a family member.

Responsible revision

The public plaque keeps the historical description and adds the three names and the provenance of the later identification. The correction is part of the record, not a silent replacement.

Still outside the evidence

The source does not publish the contributor’s name or every step of the review. We can credit the kind of evidence without inventing its missing chain of custody.

Case files opened: 1 of 3.

Correction protocol

A record can change without pretending it never changed.

  1. 01
    Preserve

    Keep the old wording, record date, source, and the exact object it described.

  2. 02
    Separate

    Distinguish a source omission from a field your own import or interface failed to show.

  3. 03
    Test

    Compare duplicates, source records, contributor evidence, and contradictions before revising.

  4. 04
    Revise

    Publish the new claim, why it changed, who or what supplied the evidence, and what remains unknown.

Sources and citation

Audit the corrections yourself.

Suggested citation: “The Archive Got It Wrong,” Nostalgia Museum, reviewed 22 August 2026.

Continue in public

Follow the evidence into another room.

Your record, when you choose

Keep the photograph and its changing knowledge together.

Preserve the Original, the back, the first caption, later family answers, and uncertainty as one private record. A correction should add provenance, not flatten the history.

Bring a photographPrivate by default. No Museum publication unless you deliberately publish elsewhere.
The Archive Got It Wrong: three photograph corrections · Nostalgia - Living Archive