How to Research Historical Fiction
Research is the part of a historical novel that feels most like progress and produces the least of it. Here is a working method: what to look up, where to look, when to stop, and how to keep nine tenths of it out of the book.
Decide how deep before you dig
A historical novel does not require a historian's knowledge. It requires the knowledge its characters have, at the resolution they have it. A dock worker in 1911 knows what a day's wages buy, which publican will run him a tab, and how cold the river is in March. He has no idea what the ships weigh or who chairs the shipping line. Research to the depth of the point of view and the book reads as researched everywhere, while you only did the work in a few places.
The practical version: list the ten or fifteen things your protagonist touches, spends, wears, eats, fears, and travels on in an ordinary week. Those are your research targets. Everything outside that list is scenery, and scenery is allowed to stay soft. It is also how you stop: open-ended reading has no natural end, and a finite list is the only thing that closes it.
Primary sources, secondary sources, and what each is for
- Primary sources were made at the time. Letters, diaries, court transcripts, newspapers, ledgers, mail-order catalogs, ship manifests, census returns, photographs, menus, maps. They give you texture and voice: what things cost, what people complained about, and the exact words they used to complain.
- Secondary sources were written later about the period. Histories, biographies, journal articles. They give you structure and causation: why the strike started, what the epidemic did to the town, which side of the argument the newspapers were on.
Use secondary sources to build the frame and primary sources to furnish the room. A history will tell you that rationing was tight in 1943. A transcribed ration book tells you how many coupons a pair of shoes cost, and that is the line a reader remembers.
An enormous amount of primary material is free and searchable. Newspaper archives such as the Library of Congress's Chronicling America. The Old Bailey Proceedings, transcribed and covering 1674 to 1913. Digitized city directories, fire insurance maps, historical mail-order catalogs. Local historical societies and museum curators answer email more often than writers expect, and specialists enjoy a precise question.
Anchor the story to the real calendar
Pin your scenes to actual dates, even if no date ever appears in the prose. Once the story sits on the real calendar, the world starts answering questions for you: what day of the week it was, what was in the morning paper, whether it was a holiday, whether the shops were shut, what the weather did.
Anchoring also catches the errors that are invisible inside a paragraph. A character cannot read about an event the day before it happens. She cannot take a train on a line that opens two years later, spend a coin that has not been minted, or telephone a house with no line to it. On the page each of those sentences looks fine; on a calendar, beside the real events, they are obvious.
Keep the chronology as a real document beside the draft, with the historical record in one column and your scenes in the other. That timeline belongs in your story bible, where it can be checked instead of remembered.
The details that actually carry a period
Readers feel a period through the small mechanics of daily life, rarely through dates. A few categories pay for the time you put into them:
- Money and wages. Prices mean little alone; the ratio is what lands. How many hours of work a coat cost, what a meal was worth against a day's pay.
- Light and heat. When does the room go dark, and what does it take to make it bright again? Gas, candles, and electricity arrived unevenly by city and by street, and the difference shapes when people worked, ate, and went to bed.
- Distance and time. How long the journey takes, how it is booked, how it smells, and what happens if you miss it.
- Work and the body. What people's hands looked like, what hurt them, what killed them, and at what age. Medicine and labor date a scene faster than architecture.
- Language and idiom. Word choice betrays a period quicker than any prop. Dated citations in a good historical dictionary will tell you when a word entered use, and a corpus tool like the Google Books Ngram Viewer gives a rough sense of when it became common, though its early data is noisy enough to treat as a hint.
One well-chosen detail outperforms five general ones. The specific coin, the specific smell, the brand of soap on the specific shelf.
How to avoid the research dump
The dump has a recognizable shape: characters explaining their own world to each other, a narrator pausing to deliver a paragraph nobody asked for, a tour of a room that contains no scene. It happens because the research was expensive and the writer wants credit for it. Four habits keep it out:
- Keep the consequence, cut the explanation. The reader can do without a paragraph on how a telegraph works. What they need is the character arriving too late because the message went by wire and the office shut at six.
- Let detail arrive as friction. Period facts land hardest when they cost somebody something: the fare that is beyond her, the stays that make stairs slow, the letter that takes nine days.
- Never let a character explain what they already know. If two people in the room both understand it, they will not say it aloud, and a reader hears the seams when they do.
- Expect a brutal ratio. Most of what you learn stays out. The research shows up as confidence in the sentences that survive, which is the only place it was ever going to show.
Verify what you wrote, not what you read
A gap opens between research and draft. You read a good source in March, wrote the scene in July, and what reached the page was your memory of the source. Memory is exactly where anachronisms live, because a wrong detail starts to feel period-correct after a few months sitting in your head unchallenged.
So run a separate pass over the finished draft, reading for claims rather than for prose. Mark every assertion about the real world: dates, prices, technologies, laws, distances, titles, the day of the week. Then confirm each one against a source you can point to, and file the source beside the claim so you can answer an editor without doing the work twice. Our guide to fact-checking a novel covers that pass in detail.
Check hardest on the details you are proudest of. The vivid fact you found early and love is the one you are least likely to re-examine, and the one a specialist reader will go straight to.
How Skarvia helps
Skarvia gives a solo writer the staff a studio would have, and research is where that staff earns its keep first. Highlight a passage in your draft, ask the question in your own words ("what did a third-class Atlantic ticket cost in 1904?"), and you get a cited answer filed into the project's Research library. The sources stay with the book instead of in a browser tab you closed in March.
When the chapter is done, run Check Facts over it. Skarvia extracts the checkable real-world claims, verifies each against retrieved evidence, and puts every finding in front of a second reviewer, the Skeptic, whose only job is to tear it apart before you ever see it. What comes back is a punch list with sources, or one quiet line: "No issues found." There is no praise channel and no rewrite button, because the bright line holds: Skarvia never writes a word for you. The period is yours to build. We just make sure it holds up.