What the Archive Leaves Out

Historians depend on documents, and it is easy to forget that a document exists only because someone made it and someone else kept it. The record of any period is not a sample drawn at random from what people did and said. It is the residue of two filters, and a historian who ignores them risks mistaking the shape of the filters for the shape of the past.

The first filter determines who writes. In most societies, writing has required literacy, leisure, and access to materials, so that those who left records tended to be those with money, office, or education. The second filter determines what is kept. Institutions preserve what serves their purposes: deeds, accounts, decrees, correspondence of the powerful. The letters of an ordinary household, whose value was not obvious to anyone but its members, are far more likely to have been thrown out. The resulting archive over-represents the literate, the official and the propertied, and it does so without any conspiracy, simply because of how records are made and saved.

A common response is to seek out the voices of those who are missing, reading the archive against its own grain. Court proceedings, tax rolls and inventories mention ordinary people incidentally, and a patient reader can learn a good deal from them. But the limits of this method deserve emphasis. These records were produced by the officials who judged, taxed and counted, for purposes that were theirs and not the subjects’. A trial record may preserve the words of an accused person, but the words were spoken under pressure and written down in a clerk’s phrasing. The person appears, but as an object of the record, shaped by the reasons it was written.

It would be a mistake to respond by imagining that beneath the official record lies an unmediated voice waiting to be recovered. The letters and diaries of ordinary people are also shaped, by the conventions of their kind, by the expectations of their readers, and by what a writer thought could be said. No source gives direct access to experience. The difference between a bishop’s letter and a servant’s is not that one is mediated and the other is not, but that the mediations differ, and that the historian knows more about some than others.

What the historian can do is to ask of every source who made it, for whom, and why it survived, and to state plainly what cannot be known. Silences, too, can be evidence. If a town’s records for a century never mention a particular group, that fact tells us something about who was being counted. But arguments from silence are treacherous, since the absence of a record does not show the absence of the thing it would have recorded. The discipline’s honest course is to grade its confidence: to say what the evidence establishes, what it suggests, and what lies beyond its reach.