Common Types of Errors in Historical Records

Historical records are never neutral. Every document, inscription, or digital file reflects the conditions of its creation—and those conditions introduce vulnerabilities to error. Understanding the categories of mistakes most frequently encountered allows researchers to approach sources with the right level of skepticism. Below are the main types of errors, each with concrete examples drawn from real historical scholarship and expanded with additional nuance.

Typographical and Scribal Errors

Even the most careful scribes make slips. In medieval manuscript culture, errors arose from tired eyes, poor lighting, or noisy scriptoria. Scribes might accidentally skip a line (haplography), repeat a phrase (dittography), or misread an abbreviation. A famous example is the so-called "Wicked Bible" of 1631, where a printer omitted the word "not" from the seventh commandment, producing "Thou shalt commit adultery." When digitizing records, OCR (optical character recognition) introduces its own typographical glitches—turning “1783” into “1789” or “George” into “Georgee.” Such errors are often subtle and require cross-checking with original images or alternative transcriptions. The scale of OCR errors can be staggering: a 2020 study of historical newspaper databases found that 20–30% of names were misread in at least one digitized edition. Researchers should always verify critical numeric and nominal data against the source image.

Transcription and Translation Errors

Copying a record by hand or translating it into another language introduces layers of potential distortion. For example, a census enumerator in 19th-century America might mishear an accent and record “Schmidt” as “Smith.” Translators working with unfamiliar idioms can change meaning: the Greek word historia once referred loosely to “inquiry,” but later translators narrowed it to “history” in the modern sense. These shifts compound when records pass through multiple hands. A notable case is the translation of the Aztec codices by Spanish friars; Nahuatl terms for ritual objects were often replaced with European equivalents, obscuring indigenous meanings. Always consult the original language version if possible, and note any discrepancies between transcriptions. For genealogical research, the FamilySearch wiki offers guides on common translation pitfalls in parish registers from various countries.

Bias and Subjectivity

No chronicler writes without a viewpoint. Political, religious, or personal biases color the selection and framing of events. A royal scribe under a king might exaggerate military victories and omit defeats. Colonial officials described indigenous cultures through an ethnocentric lens, labeling rituals as “savage.” Even court records, though legal documents, reflect the biases of the judges and clerks. The 1692 Salem witch trials exemplify how religious hysteria, personal vendettas, and legal procedure combined to produce records that demand careful contextual reading. To spot subjective distortion, read against the grain: ask whose voices are missing, and compare accounts from opposing sides. The U.S. National Archives offers guidance on evaluating primary sources for bias, including checklists for identifying loaded language and unstated assumptions.

Omission and Censorship

Errors of omission are perhaps the hardest to detect because the missing piece leaves no trace. Governments may suppress records of protest movements; private letters might be destroyed to protect reputations. In digitization projects, cost constraints lead to selective scanning—only the “most important” files are saved, creating a digital silence. The Soviet Union's practice of purging historical archives—removing photographs of disgraced officials, altering encyclopedia entries—demonstrates how systematic omission can rewrite history. Researchers should be alert to gaps in archival holdings. When a series of records stops abruptly without explanation, probe for omissions. Cross-reference with indexes, catalogues, or other collections. The Library of Congress collections often include finding aids that flag known gaps, and they maintain a "silences" section in some collection guides that explicitly discuss what is missing and why.

Errors of Dating and Chronology

Dates get confused in several ways: switching between Julian and Gregorian calendars; misreading Roman numerals; or assuming a single dating system across regions. For example, before 1752, England and its colonies used the Julian calendar while much of Europe had adopted the Gregorian—creating a discrepancy of ten days. Letters written in February 1700 (old style) might be misdated as 1699 by modern catalogers. Similarly, the French Revolutionary calendar (1793–1805) is often misinterpreted; dates like "10 Nivôse An II" require careful conversion. Always verify the calendar system in use at the time and place of the record. The Time and Date website provides conversion tools and historical context for calendar transitions across Europe.

The Problem of Forgeries and Fabrications

Beyond honest mistakes, deliberate forgeries pose a unique challenge. Medieval monks fabricated charters to claim land rights; Renaissance humanists created "lost" classical texts; modern forgers have produced diaries of historical figures. The "Hitler Diaries" scandal of 1983 showed how easily forged documents can deceive experts when desire for a sensational discovery outweighs skepticism. To detect forgeries, examine physical evidence (paper, ink, seals), linguistic anachronisms (words or phrases not in use at the time), and provenance gaps. Digital forgeries—altered images, AI-generated documents—are a growing concern. Researchers should insist on access to original physical items or high-resolution digital surrogates for critical authentication.

Strategies for Spotting Errors

Source Criticism: The Four Cs

A foundational method in historical scholarship is source criticism, traditionally evaluated on four axes: consistency, content, context, and credibility. Consistency checks whether internal details (dates, names, locations) align with each other. Content examines the plausibility of the claims—is a number improbably round? Were the weapons described actually available? Context situates the source in its production environment: who wrote it, for whom, and why. A private diary intended for no audience is generally less self-censored than a published memoir. Credibility assesses the author's expertise and access to information; a general's battle report carries different weight than a civilian's secondhand account. Applying these four lenses systematically reduces the chance of accepting flawed records.

Triangulation of Sources

No single source is infallible. Triangulation involves comparing three or more independent accounts of the same event. If two contemporary newspapers report a speech with different wording, the original transcript (if it survives) is the primary check. For early modern shipwrecks, compare official logs, court depositions from survivors, and merchant company records. Discrepancies highlight potential errors—but also reveal different perspectives. A classic case is the Battle of Little Bighorn: U.S. Army reports conflict with Lakota oral traditions, and both differ from archaeological evidence. Triangulation does not always resolve the truth, but it exposes the range of possible interpretations. The Teaching History website provides classroom exercises on triangulating primary sources, including worksheets for comparing narratives of the same event.

Digital Tools for Error Detection

Modern technology accelerates error spotting. Text comparison software like Juxta Commons or CollateX can align multiple transcriptions and flag variants. For numerical data, statistical outlier detection can flag census entries that deviate from expected distributions. WorldCat and other union catalogues let you compare metadata across libraries, catching misattributed dates or authors. However, rely on tools only as a first pass; human judgment is essential for interpreting what the tool flags as an “error.” Machine learning models trained on historical corpora can now detect likely OCR misreadings by context—for example, suggesting that "Georgee" should be "George" when neighboring text contains "Washington." The OCR-D project in Germany provides open-source tools for automatically correcting historical OCR outputs using language models.

Reading for Anomalies

Develop a habit of reading against the expected flow. Look for details that seem anachronistic—a 16th-century text using a word not coined until the 19th century. Question numbers that appear too neat: “exactly 100 soldiers died” sounds like a round-up. Check proper names against known spellings from the period. Gazetteers and biographical dictionaries can verify place names and individuals. Researchers should also note emotional language; phrases like “the vile mob” signal a bias that may distort factual reporting. Pay attention to handwriting style as well—cross-referencing signatures across documents can reveal forgery. The "Documenting the American South" project at the University of North Carolina provides side-by-side comparisons of original manuscripts and transcribed texts, allowing users to spot transcription errors directly.

Techniques for Correcting Errors

Annotating the Record Transparently

When you identify a mistake, never alter the original source. Instead, annotate your copy or dataset. In a scholarly transcription, use square brackets for inserted corrections: “The battle began at [10] AM [actually 11 AM according to the colonel’s report].” For digital humanities projects, adopt a versioning system—keep the raw OCR text, the corrected transcription, and a log of changes. Platforms like FromThePage allow collaborative transcription with built-in error logs and verified correction workflows. Always preserve the original errors as part of the historical record; they themselves can reveal something about the circumstances of creation.

Cross-Referencing with Authoritative Sources

Confirm facts using the most reliable primary sources available. For events before 1900, consult government archives, church registers, or original manuscripts rather than compiled histories. For more recent records, official statistical bureaus (e.g., the U.S. Census Bureau) provide baseline data. When a correction is needed, cite the authoritative source that supports the new information. This practice not only repairs the error but also builds trust in your own work. For example, if a transcribed colonial census records the year as "1682" but the original shows "1683" (the "2" being a misread brown spot), provide a high-resolution image snippet and a citation to the archival source. Tools like Zotero can manage these citations across large projects.

Documenting the Correction Process

Maintain a correction log for every project. List the error, the original text, the corrected version, the source of verification, and the date of correction. This transparency is crucial when the corrected record is used by others. In archival descriptions, some institutions now include “correction statements” that explain why an earlier catalog entry was revised. For example, a record might say: “Note: This letter was previously attributed to John Adams, but handwriting analysis has identified the true author as Thomas Jefferson. See NARA document 1234.” Several digital repositories, including the Europeana platform, now allow users to submit correction flags that are reviewed by curators, and the revision history is published openly.

Digital Humanities Methods for Large-Scale Correction

When dealing with thousands of records, automated correction must be guided by human oversight. Machine learning models can be trained to detect common errors—such as “0” confused with “O” in OCR—but the training data must come from hand-corrected samples. Crowdsourced correction platforms (e.g., Wikisource, Transcribe Bentham) allow volunteers to fix transcription errors, with multiple validators per page. Always keep the original version accessible alongside the corrected one so that researchers can evaluate the correction for themselves. The HathiTrust Research Center offers tools for large-scale comparison of digitized texts, enabling libraries to identify and fix errors across entire collections.

Best Practices for Ensuring History’s Reliability

Using Multiple Sources as a Default

Never rely on a single record for a factual claim. Even an “original” document may contain errors or deliberate falsehoods. Build a habit of locating at least two independent witnesses for any historical fact you intend to use in teaching or publication. For major events, aim for three to five. This not only catches errors but also enriches the narrative by revealing different viewpoints. For example, the study of the Haitian Revolution draws from French military dispatches, British naval logs, and Haitian oral traditions—each source type corrects biases in the others.

Teaching Historical Literacy

Educators play a key role in fostering error awareness. Assign students exercises that present two conflicting sources—a diary and a newspaper report—and ask them to identify the likely errors or biases. Teach the “five Ws” evaluation: Who wrote it? When and where? Why? With what authority? For what audience? Expand this with the "two additional Ws": What is missing? What is the physical context? Resources like the Library of Congress Primary Source Analysis tool provide structured worksheets that encourage critical questioning and include prompts for detecting omission and bias.

Collaboration Between Historians and Archivists

Archivists are the first line of defense against error. They establish provenance, create finding aids, and sometimes flag known inaccuracies. Historians should work closely with archival staff to understand the context of records. When you discover an error, report it to the holding institution so that catalogue entries can be updated. Many archives now have “correction” or “feedback” forms to improve metadata accuracy over time. The Society of American Archivists provides guidelines for collaborative error reporting that respect both the researcher's need for accuracy and the archivist's curation standards.

Embracing Uncertainty

Not all errors can be corrected. Some records are so fragmentary that they yield only probabilities, not certainties. In such cases, the best practice is to present the evidence transparently, note the degree of confidence, and let readers draw their own conclusions. Historical reliability improves not through the illusion of perfect records, but through honest acknowledgment of what we know and what we do not. This approach fosters a more resilient and trustworthy historical discipline. When writing or teaching, use phrases like "the surviving evidence suggests" rather than "historical records prove."

Regularly Updating Historical Databases

Digital historical resources—like census indexes, newspaper archives, and biographical databases—are never finished. They require ongoing maintenance to correct errors uncovered by researchers. Institutions that host such databases should provide clear channels for submitting corrections. Users, in turn, should check the “revision date” of an entry to see how current the information is. A database updated quarterly will be more reliable than one updated a decade ago. The FamilySearch genealogy platform uses community-driven corrections that are reviewed and rolled into the master index, with each correction traceable to the submitting researcher.

Preventive Archival Practice

Error correction is reactive; better yet is prevention. Archives and libraries can reduce future errors by implementing best practices during digitization: using high-resolution cameras, proper lighting, multiple OCR passes, and metadata standards like Dublin Core. For born-digital records, implementing checksums and version control prevents silent data corruption. Researchers can contribute by advocating for funding that prioritizes quality over speed in digitization projects. The National Digital Information Infrastructure and Preservation Program (NDIIPP) offers guidelines for sustainable digitization that minimizes the introduction of new errors.

By integrating these strategies into everyday research and teaching, historians, educators, and students can transform errors from obstacles into learning opportunities. The goal is not a flawless record—that is impossible—but a rigorous, transparent practice that makes the past as accurate as humanly possible. Every identified and corrected error strengthens the foundation of historical knowledge, allowing future generations to build upon a more trustworthy past.