Box 14 of the collection is labelled, in a typed strip glued to the front in what I would guess is the middle 1950s, MISCELLANEOUS — WOMEN'S AUXILIARY. It contains, among other things, four years of the organisation's incoming financial correspondence, including two letters that establish where the money for a building came from and who declined to give it.
That is not miscellaneous, and it is not the auxiliary's, and it took me eleven years of working in that collection to open the box, because I had read the finding aid.
The finding aid is a document historians treat as furniture. It is the guide to a collection — what is in which box, arranged by series, with a scope note — and it is produced once, usually by whoever accessioned the material, and then it is the collection, functionally, forever. Everything anybody finds in that repository is found through it. Nothing outside its categories is looked for, because looking for something you cannot name means opening every box, and nobody has that year.
So the categories that a particular archivist chose in a particular decade continue to determine what gets written about a hundred years of a town's history, and the archivist is invisible in every citation.
I do not think this is a scandal and I want to say what I think it actually is, because the sharp version of this argument leads somewhere silly.
Arrangement is not optional. A collection that arrives as ninety cartons in a barn has to be put in an order, and any order is a set of decisions about what goes with what, and each such decision is a claim about the material. The alternative to a finding aid with a bias in it is not a neutral finding aid. It is an unusable collection, and there are plenty of those, and they are worse. The archivist who typed that label was doing skilled work fast on a collection nobody was paying her for, and she got a great deal of it right.
What I want to argue is narrower: that the finding aid is a historical source in its own right, and should be read as one, and almost nobody reads it that way.
Read as a source, ours says something quite specific. The organisation's records are arranged by committee, and the committees with their own series are the ones that produced minutes. The auxiliary produced no minutes — it produced correspondence, receipts, and lists of who brought what — and so its material was not a series, it was a residue, and residue goes in miscellaneous. The arrangement did not decide the auxiliary was unimportant. It decided that minutes are the shape of a record, which is a claim about what an organisation is, and the auxiliary's material fell outside it and kept the money in it.
Now, the obvious objection, and it is arriving fast enough that I should meet it. Full-text digitisation dissolves this. Scan everything, index it, search it, and the finding aid becomes a convenience rather than a gate.
Partly. And I am for it, and I use it daily. But digitisation inherits the arrangement in two ways that are not going away. Boxes are scanned in the order somebody boxed them, and priority is set by what the finding aid says is significant, so the miscellaneous boxes go last and in many collections do not go at all. And the item-level metadata that makes a scan findable is generated from the finding aid, because there is no other source for it — which means a scanned page from box 14 is discoverable as auxiliary miscellany, and a full-text search for the name of the man who declined will find it, and a search for anything about how the building was funded will not.
The tidy conclusion available here is that we should reprocess old collections, and I will not make it, because reprocessing is enormously expensive, the people who would do it are the people currently keeping the lights on, and I have watched two institutions announce it and do a quarter of it.
The cheap thing, which I would settle for, is a habit: cite the finding aid. Put its date and its compiler in the footnote alongside the box. Do that for twenty years and the field acquires, almost for free, a record of which readings rest on which arrangements — and somebody looking at a cluster of articles that all cite the same 1950s aid will at least be able to ask the question.
I have not been able to establish who typed that label. The repository's own accession records for that decade are, as it happens, in a box marked administrative miscellaneous, and I have been through it twice.
