Sort through a box of family papers and you will meet two kinds of evidence within a minute: letters written from a training camp in 1917, and a typed family history compiled by a cousin in the 1980s. Both deserve keeping. They do not deserve to be trusted the same way, and knowing why is most of what historians mean by source criticism. The primary-versus-secondary distinction looks like a rule you learn once and forget. In practice it is a judgement you remake for every question you ask.

The Difference Is Distance, Not Format

Strip away the vocabulary and the test is simple: how far does this thing sit from the events it describes? A primary source comes from inside the period — made by people who took part, witnessed, or were caught up in what happened, or by the institutions doing the recording. A secondary source comes afterwards and interprets those traces: a monograph, a textbook chapter, a museum label, a documentary narration.

Format tells you almost nothing. A printed pamphlet from 1789 is primary. A letter written in 1950 recounting a grandmother's memories of 1789 is primary evidence of how that family remembered the Revolution, and a weak secondary source for the Revolution itself. Diaries, inventories, court records, posters, maps, ships' logs, and the unglamorous paperwork of administration all belong to the first group.

The same document, two answers

This is where beginners trip. A newspaper report from 1916 is a primary source if you are studying wartime censorship or what readers were told, and effectively a secondary source if you are trying to establish what happened at the front, because the reporter worked from dispatches, rumours, and an editor's priorities. Neither label is wrong; the question changed. Write your question at the top of the page before you classify anything, because the checklist answers only make sense against it.

A Checklist for Placing a Source

When a document lands in front of you, work through these in order. Most of the work happens in the first three.

• Who produced it, and in what capacity? A clerk, an officer, a parish priest, a shop steward — the role tells you what the person was in a position to see.
• When was it made, relative to the events? Same day, same month, or twenty years later. Memory decays and reshapes; so does paperwork.
• Participant or reporter of other people's accounts? Eyewitness testimony, hearsay, and synthesis are three different things and should not be filed together.
• What was it for, and who was meant to read it? A private diary, a public notice, and a tax return are all shaped by their purpose, just in different directions.
• Does it interpret other sources? If the maker read the evidence and drew conclusions, you are looking at secondary work, however old it is.
• Could it be both? Say so plainly. "Primary for X, secondary for Y" is a precise answer, not a hedge.

Grey Zones Worth Recognising

Memoirs and oral histories

They are primary evidence of experience and of memory, which are not the same thing. A memoir written forty years later is valuable for how the author frames a decision, what they choose to admit, and which details stay vivid. It is unreliable for dates and numbers, and it often absorbs other people's accounts without saying so. Record the interview date and treat the timeline with care.

Textbooks and encyclopedias

Firmly secondary, and useful for exactly what secondary sources are good for: orientation, vocabulary, and a map of the debate. They are also the worst place to stop. Their real value is the bibliography, which points toward the primary material and the specialist arguments underneath.

Documentaries and popular histories

Often a blend. The archive footage is a set of primary fragments; the narration, the recreations, and the dramatic music are interpretation, and sometimes invention. Separate the ingredients instead of judging the whole as one item.

Digitised archives

A scan is a copy of a copy of an object, and the catalogue record is a source in its own right. Note the collection the item came from and whether you are seeing the whole document or a selected page. Letters preserved by one proud family are a different sample of the past from a folder kept by a government department.

Reliability Is a Separate Question

None of the above tells you whether a source is trustworthy. Primary does not mean true; a forged deed is still a primary source, and an honest secondary summary can be more accurate than a self-serving diary. Judge reliability on its own axis:

1. Provenance. Can you trace the item from maker to archive? Gaps and sudden appearances matter.
2. Purpose and genre. What was the document designed to do — persuade, record, bill, comfort?
3. Access. Was the author in a position to know, or repeating what they were told?
4. Internal consistency. Do the dates, names, and figures hold together across the document?
5. Corroboration. Do two sources agree when they are genuinely independent — different people, places, and incentives?
6. Silences. What is conspicuously absent, and who had a reason for the absence?
7. Edition and translation. Are you reading the original, a Victorian translation, or an abridged version with an introduction quietly doing work?

Reading Against the Grain

Documents are built to do a job, and the job shapes what they record. A parish register counts burials without explaining them. An administrator's report describes a region through the categories his office used, and those categories are the interesting part. Ask what a record was never designed to capture, then look for sources that fill the gap — court testimony, songs, account books, photographs taken by someone with no official purpose.

Watch the vocabulary, too. "Riot" and "protest", "settler" and "occupier", "rebel" and "patriot" carry an allegiance in a single word. Copy the original phrasing into your notes verbatim, in quotation marks, before you paraphrase. You will want it later, and paraphrase quietly launders a source's bias into your own voice.

A short worked example. A soldier's letter from 1917 says the food is decent; the envelope carries a censor's stamp; the unit's supply ledger records short weights. The letter is primary evidence of what he chose to tell his family under censorship, the stamp is primary evidence of the system, and the ledger is primary evidence of supply. Together they suggest a gap between experience, permitted speech, and administrative fact. That gap is a finding, not a contradiction to be tidied away.

A Five-Minute Routine Before You Cite Anything

Write your question at the top of the page. Label each source as primary or secondary for that question, adding the qualifying phrase where it cuts both ways. Note who made it, when, for whom, and whether they could know. Then find one genuinely independent source bearing on the same point. If you cannot find a second one, say so in your writing: "only one account survives" is honest, useful scholarship, and far better than a sentence implying a consensus you have not checked.

Photo: Black_Sea / Pixabay