From Source Notes to a Historical Claim

Good source analysis can still produce a weak essay if the notes remain a stack of document summaries. Historical writing requires a transformation: from “Source A says…” to a claim about a pattern, change, relationship, or explanation that no single source supplies on its own.

The basic architecture is:

Claim because reasons, supported by evidence, connected through a warrant, limited by qualifications, and tested against alternatives.

The Craft of Research treats question formation, source evaluation, argument, audience, and research ethics as connected parts of inquiry ¹. Here you will turn those parts into a compact historical dossier.

Start with a question that evidence can change

A topic is not a question. “Women’s factory work during the war” names an area. Better questions specify a relationship:

• How did hiring categories change between 1941 and 1945?
• Which jobs opened, and which remained segregated?
• How did workers and managers describe skill differently?
• What happened to wartime gains after demobilization?

Run the change test: What possible evidence would make you answer differently? If no discovery could change the answer, you have a position to defend, not an inquiry.

Run the scale test: Can the sources speak at the level of the question? One factory newsletter cannot establish national experience. It can illuminate that factory’s official rhetoric and suggest a comparison.

Sort notes by subclaim, not source

Create evidence cards with five fields:

• subclaim: the exact proposition this card bears on;
• source + locator: stable identifier and page, frame, timestamp, or image region;
• observation: quotation, measurement, or formal feature;
• source condition: provenance, purpose, competence, and limit;
• interpretive use: support, conflict, qualification, or context.

Then group cards under subclaims. This changes the drafting unit. Instead of writing one paragraph per document, you place unlike sources into a conversation about one proposition.

Color-code observations and inferences. If your notes say “workers resented the policy,” ask whether that is a quotation, a pattern across complaints, a later researcher’s conclusion, or your inference from behavior. The National Archives’ staged analysis helps keep observation and sense-making separate before the source becomes evidence ².

Build the claim at the right strength

Use an evidence-to-language scale:

Authority does not automatically raise the verb. An official record can establish what an agency recorded, not necessarily what happened outside its categories.

Avoid universal subjects unless you have representative evidence. Replace “people believed” with “the twelve surviving letters in this collection frame…” That narrower sentence is often more meaningful because the reader can see the evidentiary boundary.

Make the warrant explicit

Evidence does not speak directly to a claim. A warrant explains the connection.

Evidence: meeting minutes repeatedly move tenant complaints to closed session.

Claim: the council treated tenant testimony as administratively sensitive.

Warrant: an institution’s repeated procedural classification indicates how it managed a category of speech.

Limit: minutes record official procedure, not every council member’s private motive or the full content of closed discussions.

Ask, “Why does this evidence count?” If the answer is merely “because the source says so,” return to provenance and competence.

Handle quotations as objects

Use the frame–quote–analyze–limit sequence:

1. Frame: identify speaker, occasion, genre, and relevance.
2. Quote: include only the language you analyze.
3. Analyze: identify wording, category, contradiction, or relation.
4. Limit: state what the quotation cannot establish alone.

Example:

In a budget hearing intended to justify the program, the director called relocations “voluntary adjustments.” The administrative phrase replaces households with an impersonal process, evidence of the agency’s public framing rather than proof that residents experienced the moves as voluntary.

The source condition belongs inside the analysis, not in a detached disclaimer.

Distinguish sequence, mechanism, and cause

“A happened before B” is chronology. “A caused B” requires more.

For a causal claim, document:

• sequence: cause precedes outcome;
• mechanism: a plausible process connects them;
• covariation or contrast: outcome changes where or when the cause changes, when evidence permits;
• alternatives: other plausible causes are considered;
• contemporary linkage: actors or records connect the events, evaluated as situated evidence;
• scope: where, when, and for whom the relation holds.

Historical evidence rarely behaves like a controlled experiment. That makes calibration more important, not less. You can argue that a policy “contributed to,” “accelerated,” “redirected,” or “made possible” an outcome without claiming it was the sole cause.

Cite the item you actually used

For archival or digitized material, record:

• creator and title or descriptive name;
• date;
• collection, box/folder or call number when available;
• repository;
• stable item URL;
• version information: scan, transcript, translation, edition, crop;
• access date if the citation style requires it.

Keep an image’s catalog page with the downloaded file. A direct JPEG URL proves location, not identity, rights, or context. For a quotation, store a page or frame locator at the moment of note-taking.

Paraphrase from memory after closing the source, then reopen it to verify accuracy. This creates language independent of the source’s syntax while exposing misunderstanding. Use quotation marks during note-taking, not only in the final draft.

Build a one-page evidence dossier

Your module project contains:

1. Question: one sentence, bounded by time, place, and problem.
2. Working answer: 40–60 words with a confidence term.
3. Three subclaims: each with at least two evidence cards.
4. Source map: production chains and dependencies.
5. Counterevidence: the strongest conflicting item or pattern.
6. Silence note: what the archive cannot show and why.
7. Revision note: how the claim changed during sourcing.
8. Next-source request: the exact record that would most improve confidence.

Then draft one 350-word argument using at least three sources from two kinds of media. Each paragraph should contain a relationship among sources, not a tour of them.

Audit the argument

Underline claims, circle evidence, box warrants, and bracket qualifications. If a paragraph has evidence but no warrant, explain the connection. If it has claims but no evidence, narrow or support them. If every sentence is qualified, decide what the evidence positively establishes.

Historical thinking is not generalized doubt. It is confidence in proportion to source capacity, independence, context, and comparison ³.

Exit statement

Complete:

The available evidence most strongly supports . I use the term ___ rather than ___ because . The claim applies to ___ and should not be generalized to . The next source most likely to change my answer is ___ because .

If you can fill every blank precisely, your source notes have become a historical argument.