Nobody Ever Has Good Sources
The fantasy is a document that answers the question. The reality is four documents that partly contradict each other, all written by people with reasons, none of them addressing what you actually want to know. Source criticism is not the stage before the argument. It is the argument.
The Questions, Every Time
- What is it? Genre first — a chronicle, a charter, a letter, a report. Genre sets the conventions and therefore the expected distortions.
- Who made it, when, and how far from the event? An eyewitness the same evening is not a compiler three centuries later working from a lost source.
- Why? Every source was made for something. Find the purpose and you have found the direction of the bias.
- For whom? Audience shapes content — a private letter, a public despatch, and a report to a superior say different things about the same afternoon.
- What does it take for granted? The unwitting testimony, and usually the best part.
- What is missing, and why? Silence is evidence, but you must decide whether it is a silence of absence, of filtering, or of physics.
Corroboration Is About Independence
Three sources agreeing sounds decisive. It is worth nothing if all three copied the fourth.
Ask whether the agreement is independent. Independent sources have separate causal paths to the event; dependent ones share an ancestor and multiply nothing but the copy count. Most spectacular ancient "confirmations" are one lost source with three descendants — and the error is not that the historian trusted the sources, but that they counted them.
The one rule of thumb worth memorizing: weigh testimony against interest. When a source reports something against its own interest, that report is unusually strong — a tax record admitting an exemption, a general's despatch conceding a setback, a chronicle recording its patron's defeat. Nobody invents inconvenient facts about themselves. Conversely, a source reporting exactly what it wants to be true has told you about its wants, not the world.
Contradiction Is a Finding
When sources conflict, the amateur picks the one they like. The historian asks why they differ, because the pattern of disagreement is itself evidence:
- Do they diverge exactly where their interests diverge? Then the divergence is the finding — you have located the dispute.
- Do they agree on the mundane and differ on the significant? That is the normal shape of politically live events, and the mundane agreement is your firmest ground.
- Does one contain a detail no purpose explains? Purposeless detail is expensive to invent, and often the most trustworthy thing in the document.
Writing It Honestly
State the evidence, its provenance, and the inference. Distinguish "the record shows" from "it is likely" from "I suspect" — and give the reader enough to disagree with you. The measure of an argument is not confidence but whether a reader can see exactly where it would break.
An argument that survives because its foundations were hidden is worth less than one that fails visibly. And the honest historian's final duty is the unglamorous one: say what the evidence cannot settle, rather than filling the gap with fluency. Most of the past is not recoverable, and a discipline that admits this is more trustworthy — not less — than one that always has an answer.