There is a particular pleasure in discovering a fact that is simply too good to leave out of a book. It is surprising, colourful and exactly what the chapter needs. You can already imagine the reader pausing over it.
Then comes the less comfortable question.
Is it actually true?
For an author of non-fiction, that question should be a regular companion. Preferably one who pulls up a chair before the manuscript reaches the reader, rather than arriving afterwards with a raised eyebrow and a collection of inconvenient documents.
Research is part of the responsibility we accept when we ask people to trust our words.
A good story still needs a sound foundation
I write across subjects that invite curiosity: British history, folklore, traditions, everyday life and the peculiar things people once believed, attempted or considered perfectly sensible.
These subjects offer wonderful material. They also require care.
An entertaining account can travel a considerable distance from its original source. A tentative suggestion becomes a confident statement. A local story acquires a precise date. An explanation that sounds plausible becomes the explanation everyone repeats.
By the time an author encounters it, the account may look thoroughly established. Yet repetition alone does not tell us whether it is accurate.
Five websites repeating the same sentence may all depend on one earlier article. That is one trail to investigate, rather than five independent confirmations.
A lively writing style can carry a reader through complicated material. It cannot turn an unsupported claim into a supported one.
Start with the question, not the conclusion
There is a temptation to begin research with an answer already in mind.
You want a custom to be ancient. You want an incident to demonstrate extraordinary courage. You want an invention to have changed everything overnight.
The trouble begins when you collect only material that agrees with the story you hoped to tell.
A better starting point is a question. What evidence survives? When does the account first appear? Who recorded it? What can that source reasonably establish?
Sometimes the research strengthens your original idea. Sometimes it complicates it. Occasionally, it removes the best anecdote in the chapter.
That last possibility is irritating. It is also part of the job.
The reader deserves the account the evidence supports, even when it requires a less convenient paragraph.
Go beyond the first search result
An online search can help you find material, but finding a statement and checking a statement are different tasks.
Where possible, follow the trail towards the underlying evidence. That might mean a contemporary newspaper, a letter, an official record, an interview, a museum catalogue or a carefully researched study.
Then ask what you are looking at.
Was the writer present? Were they recording an event at the time or remembering it many years later? Was the document intended to inform, persuade, entertain or defend somebody’s reputation?
Even a contemporary source needs interpretation. A newspaper report can contain errors. A witness has a limited view. A personal account may be sincere while leaving important details uncertain.
“Primary source” is a useful description of a source’s relationship to the subject. It is not a certificate of perfection.
Good research means considering evidence in context and comparing it with other relevant material.
History, folklore and fiction need clear boundaries
For anyone writing about folklore, this distinction matters enormously.
A village legend can be worth recording without being a verified historical event. A ghost story may reveal something about a place, its identity or its storytelling traditions without proving that a spectral coach ever negotiated the high street.
The wording should make that clear.
“Local legend tells of…” makes a different promise from “This happened…”
Likewise, an account documented in the nineteenth century does not, by itself, establish that the same story circulated a thousand years earlier. If an earlier origin is suggested, explain the basis for that suggestion and its limitations.
There is plenty of room for mystery in non-fiction. There is also room for disagreement, incomplete records and stories whose beginnings remain elusive.
We do not have to settle every uncertainty to write something worthwhile. We do have to let readers see where the uncertainty lies.
Check the small details
Research is not only about the grand argument. It is also about the apparently minor detail that slips past because it feels familiar.
A name. A date. A place. A military rank. The spelling of a village. The difference between an event being proposed, authorised and actually carried out.
If you include a quotation, check its wording and attribution against a reliable source. If you use a number, establish what it measures, which period it covers and where it came from.
Words such as “first”, “only”, “always” and “never” deserve particular attention. They make substantial claims in very little space.
An author might reasonably establish that something is the earliest example they have found. Claiming it is the first example ever requires rather more confidence.
Accuracy often improves when we become more precise about what we can demonstrate.
Keep notes you can follow later
There is a special kind of frustration reserved for the author who remembers finding an excellent source but has no idea where.
Somewhere online. Possibly a museum. There was a photograph. It may have been Tuesday.
That is not a reference system.
Record the author or organisation, the title, the publication date where available, the page number or relevant section, and the web address for online material. Record when you accessed a page if that will help you identify the version you used.
Keep copied quotations clearly separate from your own notes and draft wording. Otherwise, someone else’s sentence can find its way into your manuscript disguised as something you thought of over breakfast.
Useful notes should also record doubts. Mark the claim that needs checking. Keep the conflicting date. Preserve the question you have not answered.
Your future self will be grateful.
Use AI with an author’s judgement
AI can help organise questions, suggest avenues to explore and identify passages that need clearer explanation. Its output still needs checking.
A convincing answer is not evidence. A plausible quotation needs a traceable source. A reference needs to exist and support the claim attached to it.
The responsibility remains with the person publishing the work.
If a tool helps me write faster, I still need to know whether the resulting words are accurate. Speed does not transfer responsibility elsewhere.
The same principle applies to any convenient summary: use it to help you investigate, then examine the evidence before relying on it.
Make room for “we do not know”
There is nothing embarrassing about acknowledging a genuine gap in the record.
“The evidence is limited.”
“Accounts differ.”
“The origin remains uncertain.”
These can be useful sentences when they describe the position fairly. They tell readers how much weight to place on what follows.
Non-fiction can be engaging without pretending the past left us a complete set of instructions.
Correct mistakes when you find them
Careful work can still contain mistakes. What matters next is how an author responds.
Check the issue, establish the correction and amend the work where possible. If the error substantially affects what readers have been told, make the correction clear to those readers.
A digital edition offers opportunities to update a text, but revision still needs care. Correct the passage, check any related claims and keep a record of what changed.
Being willing to correct a book is part of taking its readers seriously.
Readers deserve the work behind the words
People buy non-fiction because they want to learn, understand, investigate or see a familiar subject more clearly. They should be able to distinguish established evidence from interpretation, tradition and speculation.
They also deserve readable writing. Research does not require every sentence to arrive wearing a ceremonial gown and carrying a footnote the size of a small wardrobe.
There is room for warmth, humour, atmosphere and a good turn of phrase. The foundations simply need to hold.
For me, that is the standard worth working towards at Rob’s Books: interesting subjects, thoughtful storytelling and care over what the reader is being asked to believe.
Before publishing the remarkable fact, follow its trail. Before repeating the quotation, check it. Before declaring that everybody knows something, find out how anybody knows it.And if the best sentence in the chapter turns out to be wrong, change the sentence.
You can explore books, British history, folklore and more at Rob’s Books.
