CAUTION, NOVELISTS: historians at work - Emma Darwin

The other day I was talking to a writer friend about my current project, and she urged me to read a particular novel set in the same period, centred on the same sort of people. I made a polite but non-committal answer, because I can't tell you how certain I was that I'd do nothing of the kind.

The thing is, someone else's historical novel is their re-visioning and re-imagining of that time. How the novel is structured, how it's voiced, what historical material it uses, what it writes on the spaces between the facts, what it elaborates in the gaps in the record... all of those are the product of that writer's self. Their consciousness, their nature as a storyteller, is the creative engine and the organising principle, not mine. It's not just a question of not wanting to plagiarise unconsciously (and Heyer spotted Cartland's plagiarism in Cartland's taking things Heyer had invented as historical material). It's much more that while all writers of historical fiction are at one remove from the world they're trying to evoke, I don't want mine to be a third-hand world.

Charles Dickens' French Revolution in A Tale of Two Cities is not the same as Hilary Mantel's in A Place of Greater Safety, or Marge Piercy's contemporaneous City of Darkness, City of Light. Georgette Heyer's Waterloo is not Thackeray's (she was shocked by how wrong many of his facts are!), and I avoided both when I was writing The Mathematics of Love. From the moment when I realise I want to write about a period, I stop reading historical fiction set in it.

What's less obvious is that a similar problem exists with the history books. What a historian writes is also - to an extent that historians haven't always been willing to acknowledge - filtered through that historian's consciousness: not just their personality but the biases and discourses of their time. Carlyle's French Revolution is not the same as Simon Schama's, and both will leave things out - different things - that I need to know about.

Mind you, History, as a discipline, has changed. J H Elliott's classic short study, Imperial Spain, was first published in 1963, and has been on syllabuses and reading lists ever since. For the second, revised edition, which came out in 2002, he points out that a book of this kind written now would never have so little to say about the experience of women. In the meantime, Women's History has brought us treasure troves such as Marilyn Yalom's History of the Wife. As someone who writes battles and politics, but also childbirth and cooking, that's not the only kind of book I need, but it helps a lot.

But there's a more fundamental problem. History, as a discipline, is about finding the larger patterns and forces which shaped lives in the past. An honest historian may acknowledge some evidence which exists but has yet to be fitted in. But still, the project will be to synthesise things to explain the whole picture. And yet always, as a novelist, I'm aware that the opposite was probably also true. You know that plan of a medieval village you copied into your exercise book, aged around ten? No one village looks like that because they all have their quirks, but that plan is more true as History, beause it presents the essence of the matter. If you think round your friends I'd put money on every single one of them having several characteristics which don't fit the norm for their job/background/class/ethnicity/gender/nationality. The essence of gender history is that husbands batter wives, but that doesn't mean that no wife has ever battered her husband, and my characters are individuals, not essences. If I want to put a battered husband in my novel, I shall. I'll have to work harder to convince the reader, but that's never a bad thing for me as a writer, or for my reader. As I was talking about on my own blog a while ago, the expected thing slips past the reader too easily: it's the surprising, the off-beat, the taking-aback thing, done properly, which catches the reader and holds them long enough for the story to come alive.

Of course ordinary, individual lives aren't ignored by historians: history "from below" is big business. We've got more sophisticated, too: a book such as Judith Flanders' The Victorian House starts with the evidence of books about etiquette and household management, but then the much more complex and slippery reality needs a different kind of teasing out. Even the absence of evidence can now have shrewd insights read into it, which is something at which novelists might be able to teach historians a thing or two: apparently Hilary Mantel's Wolf Hall is on several undergraduate History syllabuses, not because it should be read as History, but as an example of the kind of imaginative effort that should be part of any historian's practice.

And you can go back to original sources. If you're trying to bring to life a real, historical character then with luck there will be plenty of those - though again, you have to be wary of the biases and discourses forming anything with more creative input than an account book or a warrant of execution. As Alison Weir describes, Eleanor of Aquitaine's contemporaries didn't find her day-to-day presence, ruling the kingdom for Henry II, worthy of record: women were only good for sex and procreation, so that's what was written down about Eleanor, and you need to go back to charters and Privy Council records to discover otherwise. Yet (as I found, similarly, when researching Elizabeth Woodville for A Secret Alchemy), later historians still didn't examine, let alone challenge, the version of the Queen that was handed down from her own time: the discourse of their times wasn't so different. What we think of as History is, much of the time, just someone else's version of it.

So original sources are no guarantee that your apprehension of your period hasn't been pre-sifted, limited, interfered with, and that's always supposing that you can find the sources. If your novel's about a fishwife in 12th century Cumberland, the original sources will all be in Welsh, except when they're in Norman French, and they're extremely unlikely to be in any such woman's own hand or voice (the same is true for, say, accounts of ordinary soldiers before the 19th century.) What's more, one particular, real fishwife's experience may not be at all what you want or can use, because your novel isn't about her, it's about someone else. So you might well be better off with more general accounts: a History of the Wife, and a History of the Medieval Cumbrian Fishing Industry, and any number of other histories of religion, food, textiles, architecture, transport, politics. And then you do the maths - the listening - the imagining - the dreaming - in the spaces between all that history. Historians are in the business of synthesising a general picture from particular experience. Novelists do it the other way round.