What Survives Is Not a Sample
The first thing to understand about historical evidence is that it is not a random sample of the past — it is a biased survivor. Three filters run before a historian ever opens a box:
- What got recorded. Literate, powerful, and institutional actors write things down. A tax assessor generates documents; a shepherd generates none.
- What got kept. Archives preserve what an institution thought worth preserving, which is what served the institution.
- What survived. Fire, damp, war, and recycling did the rest — and parchment outlasts papyrus, stone outlasts wood, gold outlasts bread.
Each filter tilts the same way: toward the wealthy, the literate, the male, the metropolitan, the recent. So the shape of the archive is a political fact, and reading it naively means reproducing the priorities of everyone who ever held power over a document.
The Argument From Silence, and Its Traps
The tempting inference — "no evidence of X, therefore no X" — is valid only when you have established that if X had happened, evidence would very likely have survived. That condition is rarely met.
The test to apply: before treating a silence as evidence, ask what the record would look like if the thing HAD happened. If the answer is "exactly the same", the silence tells you nothing about the past — it tells you about the archive.
Bias Is Not Contamination
A partisan source is not a worse source; it is a source about different things. A hostile chronicler is unreliable on his enemy's motives and excellent on what accusations were plausible enough to make. Propaganda is worthless as a description of conditions and priceless as evidence of what the regime needed people to believe.
| Question | What a biased source gives you |
|---|---|
| Did it happen? | Weak — check against independent traces |
| What was said? | Often strong |
| What was believable? | Very strong — bias reveals the plausible |
| What did the author want? | Excellent — bias is the datum |
The corollaries do real work. Testimony against interest is unusually strong: a chronicler admitting his own side's atrocity is reporting something he had every reason to hide. Incidental detail is strong: a bill of lading listing cargo has no motive to lie about the price of pepper, because it was not trying to tell you anything about the price of pepper. What a document says in passing, while arguing about something else, is where much of social history lives.
Objectivity, Properly Defined
The naive picture — the historian as a mirror, with no standpoint — is impossible, since selection requires criteria and criteria come from questions. But the collapse into "therefore all accounts are equally valid" is worse, and it does not survive one minute of contact with practice: some claims are refuted by the evidence and others are not, and no amount of standpoint-talk makes a forged document genuine.
Objectivity is a procedure, not a position. It means: state your criteria; seek disconfirming evidence rather than favorable evidence; represent your sources' arguments in their strongest form; distinguish what the evidence shows from what you infer; and publish so others can check. It is not the absence of a standpoint. It is the discipline of holding one accountable.