Sourcing when little survives

Start with the gap, not the story

When the obvious archive is empty, the honest move is to make the absence itself the first finding. Before you write a sentence of narrative, write a short inventory of what you hoped to find and what actually exists. The National Archives in the United States notes that of all federal documents and materials created, only about 1 to 3 percent are kept permanently for legal or historical reasons (https://www.archives.gov/research/start). A thin file is not a failure of your effort. It is the normal condition of the record. Once you accept that, your piece stops pretending to be complete and starts being trustworthy.

What counts as an archive when the shelves are bare?

An archive is any stable body of records you can cite and someone else can check. It can be a national catalogue, a local library holding, a company ledger, a photograph with a date, a map, a tax list, or a single letter with a legible sender. The National Archives (UK) treats its Discovery catalogue as the entry point to records and also points researchers toward research guides, an A to Z index, and subject collections covering family history, military and maritime, social and cultural history, land and maps, and more (https://www.nationalarchives.gov.uk/help-with-your-research/). Your archive may be smaller, but the test is the same: can you name it, locate it, and describe its limits? If yes, it is usable. If no, it is background reading, not evidence.

How do you widen the search without wasting weeks?

Work outward in rings. Ring one is the catalogue you already know. Ring two is the adjacent catalogue: the same institution's other collections, or a related agency. Ring three is the substitute source: a published index, a register, a later summary, a photograph collection. The UK guidance explicitly offers "reading old documents" tutorials and a "find a research guide" path, which is a signal that when direct records are hard to read or hard to find, the responsible step is to learn the finding aids first (https://www.nationalarchives.gov.uk/help-with-your-research/). Set a time box for each ring. If a ring produces nothing citable, record that it produced nothing, with the search terms you used. That record becomes a footnote, and footnotes about absence are legitimate scholarship.

Which source do you trust when two of them disagree?

Trust the one closest to the event, produced nearest in time, by someone with a reason to record it accurately. Then trust the one whose custody chain you can describe. A catalogue entry that tells you which collection a document belongs to is more useful than a summary that does not. The US National Archives describes a four step path: determine your topic, gather information about it, find records and information about records, then decide whether to visit and plan that visit (https://www.archives.gov/research/start). That middle step, finding information about records, is where you learn whether a document is likely to exist at all. Do that before you commit to a trip, a fee, or a strong claim.

How do you separate what a document says from what you think it means?

Use two columns. Left column: what the document states, in plain paraphrase, with the catalogue reference. Right column: what you infer, with your reasoning. Keep them physically separate in your draft, even if the final article blends them. The UK guidance links to a dedicated page on citing records and catalogues, which is a reminder that citation is not decoration; it is the boundary between evidence and interpretation (https://www.nationalarchives.gov.uk/help-with-your-research/). If your right column is much longer than your left, your piece is mostly inference. That can still be published, but it must be labelled as such, and the reader must be told what would change your mind.

What should the article actually contain?

A useful thin archive piece usually has five parts. First, a short statement of the question and why it matters. Second, a plain description of the record base: what exists, what does not, and how you know. Third, the evidence itself, quoted or paraphrased tightly. Fourth, the inference, marked as inference, with the strongest counter argument you can state fairly. Fifth, an open list of what a reader could check next. This structure is borrowed from archival practice, not from magazine convention, and it is what makes a small piece credible when a large one is impossible.

Decision checklist

Situation	Do this	Avoid this
Catalogue returns nothing	Record search terms and date of search	Claiming the thing never existed
Only a later summary survives	Cite the summary and name the missing original	Quoting the summary as if it were the original
Document is hard to read	Use a reading guide or tutorial before guessing	Filling gaps with plausible words
Two sources conflict	Prefer nearer in time and clearer custody	Averaging the two into a false middle
You want to visit in person	Plan the visit and confirm access rules	Assuming everything is digitised
A claim rests only on your reasoning	Mark it as interpretation	Presenting it as documented fact

How do you write the gaps without making the piece feel empty?

Name the gap early, then spend the rest of the piece doing real work inside it. Describe the shape of the surviving record. Quote the fragments you have. Explain what each fragment can and cannot support. Then tell the reader what the absence suggests: not a conspiracy, usually, but a filing decision, a lost box, a practice of not recording certain people or transactions. The honest article does not mourn the missing archive. It maps it, and then it hands the reader a method they can reuse on their own topic.

What should you do before publishing?

Check every factual sentence against a source you can name. Check that each inference is visibly marked. Check that your citations point to real catalogues and guides, not to vague institutions. For current access rules, opening hours, copying services, and any fees, consult the relevant archive's own current guidance rather than relying on an older article, including this one. If your research touches legal, medical, or financial questions, treat the archive as context only and direct readers to current official guidance. The discipline is the same at every scale: say what you found, say what you did not find, and say how you know the difference.

When is a thin archive enough to publish?

It is enough when the question is clear, the evidence is cited, the limits are stated, and a reader could repeat your search. It is not enough when the article needs a completeness the record cannot give. That is the whole test. A piece built on three verifiable documents and one honest paragraph about the missing rest will outlast a piece built on a confident story with nothing behind it.