Transparency
Corrections.
Public, dated, linked. An archive that argues with the textbooks owes the same standard of accountability it asks of them.
In short
Does Unsilenced publish its corrections?
Yes. Every substantive change to a claim, figure, date or image attribution is recorded on this page with its date and the chapter it affects, and applied to all six language editions at once. Typographical fixes are made silently. Errors can be reported at any time through the contact page.
Report an error
If you've found a wrong date, a misattributed photograph, a superseded figure, a broken citation or an outright mistake, please email us via the contact page. Include the URL of the page, the specific claim and your source. Substantive fixes are logged below within seven days; trivial typographical fixes are made silently.
Our standing rules for sourcing and figures are documented at /methodology.
What counts as a correction
Logged — substantive
A revised casualty estimate, a corrected date or place, a reattributed or replaced photograph, a citation that turns out not to support the sentence attached to it, a misspelled personal or place name, or a claim withdrawn for want of evidence.
Silent — cosmetic
Spelling, punctuation, broken links repointed to the same document, layout and translation phrasing that does not change meaning.
Not a correction — interpretation
Disagreement with the archive's framing, or a preference for a higher death toll than the lower-bound rule allows. These are answered by letter, and where the scholarship is genuinely split the dispute is described on the page itself.
How a correction moves
01 — received
A reader sends the page URL, the quoted sentence, and a source.
02 — checked against the original
The primary source or monograph is consulted directly, not a summary of it.
03 — applied in six languages
The English chapter and its Spanish, French, German, Portuguese and Arabic editions are updated together.
04 — logged and dated
The entry appears below, and the chapter's "last updated" date is bumped.
Log
2026-01-01
Site launch
Corrections log opened. All chapters published in v1.0 state.
Common questions
What counts as a correction on Unsilenced?
Any change that alters what a reader would take away from the page: a revised death-toll figure, a corrected date, a reattributed photograph, a fixed or replaced citation, a changed name or place, or a claim withdrawn for lack of support. These are dated and logged publicly. Typography, spelling and layout fixes are made silently.
Why publish a corrections log at all?
Because an archive that argues with school textbooks has to be auditable. A visible log lets a reader see how often the archive is wrong, in which direction, and how quickly it is fixed — which is the only honest way to earn trust in contested history.
How are disputed death tolls handled?
Where credible scholarly estimates disagree, the archive publishes the lower bound of the majority range and names both sides of the dispute on the page. Replacing a low figure with a higher one is therefore not usually treated as a correction unless the lower estimate has been superseded in the literature.
Are corrections applied to every language edition?
Yes. A substantive correction is applied to the English page and to the Spanish, French, German, Portuguese and Arabic editions of the same chapter, so the six versions never drift apart on facts or figures.
In short
How does Unsilenced decide what to change once a chapter is published?
Publication is not the end of the editorial process here; it is the point at which claims become checkable by strangers, which is the whole design. Every chapter carries the sources it relied on, so a correction almost always starts as a reader — or an author re-reading the underlying monograph — finding a mismatch between what the page says and what the cited source actually supports. The archive does not treat a correction as an admission that the project failed; it treats an uncorrected error, once found, as the failure.
What changes and what does not follows a fixed distinction: facts, dates, names, attributions and figures are corrected against the cited literature; interpretation, framing and the choice of which lower-bound estimate to lead with are editorial positions the archive holds and defends rather than 'corrects' on request. The corrections log exists to make the first category visible and auditable, and the methodology page exists to make the second category explicit so it is never mistaken for a factual error.
Correction classes
Six kinds of change and how each is handled
| Class of correction | What triggers it | How it is handled |
|---|---|---|
| Figure revision | A cited estimate is superseded by newer peer-reviewed work | Old figure replaced, prior estimate and its source kept visible in text |
| Date or place error | A verifiable factual mismatch with the primary record | Corrected silently if trivial, logged if it changes a claim's meaning |
| Image misattribution | A photograph's date, subject or institution is wrong | Caption and Wikimedia/archive link corrected, credit re-checked |
| Broken or mismatched citation | A footnote points to a source that does not support the sentence | Citation replaced or the sentence narrowed to what the source shows |
| Name or spelling error | A personal, ethnic or place name is misspelled or outdated | Corrected against current scholarly usage, noted if politically contested |
| Withdrawn claim | A claim cannot be traced to any credible source on review | Sentence removed rather than left unsupported; noted if previously prominent |
Categories used to triage incoming reports, not a record of specific past corrections.
Severity
Not every fix is announced the same way
A cosmetic error — a stray comma, a rendering glitch, a dead link repointed to the same document — carries no meaning risk and is fixed without a log entry, because logging it would clutter a page meant to track substance. A substantive error is different by definition: it is any change that would alter what an attentive reader takes away from the page, and those are the only entries that belong on a public corrections page.
Between those two poles sits a smaller category worth naming: corrections that are substantive but do not change the argument, such as tightening a death-toll range as a new academic edition appears. These are logged, but logged plainly, without treating a normal advance in scholarship as if it were an admission of prior negligence.
Why figures move
Numbers change because the scholarship does, not because the archive guessed
Colonial demography is a field of ongoing archival recovery: mission burial registers, tax rolls, shipping manifests and oral-history projects continue to surface, and estimates for events a century or two old are still being revised in journals today. A figure published in 1990 and a figure published in 2020 can both be scholarly and still disagree, because the underlying evidence base has grown.
When that happens here, the archive updates to the new lower-bound within the current mainstream range, cites the newer work by name, and does not pretend the earlier figure was invented rather than superseded. Readers who want the history of a number, not just its current value, can trace it through the cited sources on the chapter itself.
Not a correction
What this process deliberately does not fix
Requests to raise a death toll to the highest defensible figure, to drop the word 'genocide' or 'colony' for a softer term, or to reframe an atrocity as a product of its era are not corrections, because they are not claims about what the sources say — they are disagreements with editorial judgment the archive has made openly, on the methodology page, and stands behind.
Where scholarship is genuinely, currently split on a figure or a characterization, that split is described on the page itself rather than resolved by picking a side and calling it fact. That is a sourcing decision, and it is reviewed like one; it is not treated as an open wound waiting for a reader's preferred number to close it.
FAQ
Frequently asked
- Does a correction mean the original research was wrong?
- Not usually. Most corrections here are figures superseded by newer scholarship, images later re-identified by an archive, or citations narrowed to what a source actually supports — normal maintenance of a living reference work, not evidence the original claim was fabricated.
- Who decides whether a report becomes a logged correction?
- The editorial team checks the report against the primary source or monograph directly, not against a summary of it, before any change is made. A report that turns out to rest on a misreading of the existing source is answered but not logged as a correction.
- Can a reader dispute the framing of a chapter without it being called a correction request?
- Yes, and this happens regularly. Disagreement with how an event is characterized, or with the choice to call a territory a colony, is a substantive conversation the archive welcomes through its contact page, but it is answered as editorial dialogue, not filed as an error.
- How long does a genuine factual correction take once confirmed?
- Once the underlying source has been checked and the change confirmed, the correction is applied to all six language editions and logged, generally within about a week of confirmation — the archive does not sit on a confirmed error waiting for a convenient publishing cycle.
- Does correcting one language edition automatically fix the others?
- No — each language edition is a separately written text, not a machine translation of English, so a confirmed correction has to be applied by hand to each of the six versions. That is deliberate: it is also the point at which a translator can catch a second error the English original missed.