How to Read a Primary Source Like a Historian
2026-08-04 · 8 min read
Sourcing, contextualising and corroborating: the three habits that turn a document from a wall of old text into evidence.
History is an argument about evidence School history is often taught as a settled sequence of events to be memorised. Actual historical work is closer to a legal case: you have incomplete, biased, contradictory documents, and you build the most defensible account they support. Learning to read a primary source is learning to handle that evidence.
Researchers who study how historians read found that experts do three things novices skip almost entirely.
Sourcing: read the attribution first Before the text, read who wrote it, when, for whom, and why. A diary entry written for nobody, a speech written to persuade a hostile parliament, and a memoir written thirty years later for a reputation are three different kinds of evidence, even about the same event.
Ask what the author was in a position to know. An ambassador reporting a battle he did not attend is reporting rumour, however confident the prose. Ask what the author wanted. Bias is not a reason to discard a document — it is information about the author and their audience, and sometimes the bias is the most useful thing in the source.
Contextualising: put it back in its moment A document means what it meant then. Words shift: "liberty," "commerce," "nation," "revolution" all carried different loads in 1770 than in 2026. Assumptions shift more — about who counted as a person, what the state could legitimately do, how the natural world worked.
Ask what else was happening that month. A pamphlet published a week after a massacre is doing something different from the same argument published in a quiet year. Ask who could read it, how it circulated, and what it cost. Reach is part of meaning.
Corroborating: never trust one document Any single source can be mistaken, self-serving or fabricated. Historians build claims where independent sources converge — and pay close attention where they diverge. Contradiction is not a dead end; it is often the most interesting part, because someone had a reason to tell it differently.
Watch for false independence. Three newspapers running the same wire report are one source, not three. A textbook citing a monograph citing a single letter is still the letter.
Reading for silence What a document does not say can matter as much as what it does. Census categories reveal what a state thought worth counting. A plantation ledger recording enslaved people alongside livestock tells you about a worldview in its structure, not its sentences. Ask whose voices are missing, and whether the absence is accidental or produced.
A practical routine Work in this order and it becomes fast: read attribution, note author's position and purpose, skim for the main claim, read closely once, mark unfamiliar terms and check them, write two sentences on what this source is good evidence for and one on what it is not, then place it against another source on the same event.
That last step is the whole discipline compressed. "Good evidence for what?" is a better question than "is this true?" because most documents are reliable about some things and unreliable about others. A propaganda poster is weak evidence about battlefield conditions and excellent evidence about what a government wanted its population to believe.