In 2005, the novelist Kate Grenville published The Secret River, a book about a petty criminal transported to New South Wales who builds a new life on "virgin" land already occupied by Aboriginal people. It won the Commonwealth Writers' Prize and was shortlisted for the Man Booker Prize. Australian film director Neil Armfield described it as "a key to understanding our past".

How had Grenville managed to write so convincingly about another era? Along with a great deal of library research, she had set out alone into the bush and even used a rag dipped in lamb's fat to make a traditional "slush lamp". When she lit the rag and was engulfed by the foul smell it produced, she reported in her account of the book's genesis, Searching for the Secret River: The Story Behind the Bestselling Novel (2007), she "learned more about life in a bark hut on the Hawkesbury (River) in 1817 than all the books in the world could have told me".

Her claim is not unusual; we often hear that novelists use empathy to transport themselves to places mere historians can never reach. They are said to offer a different - and perhaps a deeper - kind of truth than those constrained by hard factual evidence. Many believed that Grenville was implying that she had created a new sort of history.

Such claims can make professional historians furious. Inga Clendinnen, who for many years lectured in history at La Trobe University in Victoria, Australia, says that Grenville's transposition of episodes to different times and places made her novel "not only not history but, in my admittedly very austere view, anti-history".

ÁñÁ«ÊÓƵ

Nor can Clendinnen relate to novelists' famed "empathy". Working on the Aztecs, her understanding suddenly hit a brick wall when dealing with newborn baby boys sacrificed on the killing stone.

"With that action," Clendinnen recalled, "a mist dropped: a mist that would only be dispersed, and then only partially, by long, cautious labour. Untutored 'subjectivity', 'empathy', was of not the least use to me inside the Aztec world."

ÁñÁ«ÊÓƵ

In the same way, if to a lesser extent, there are no intuitive shortcuts to the early 19th-century world described in Grenville's novel. Only rigorous research can give us insights into how it was.

So should there be movement on the borders between academic history and historical fiction, or is it essential that they remain tightly policed? Do novels encourage a healthy interest in the past, create myths that get in the way of more serious analysis, or simply attempt to do such different things from academic work that they are irrelevant either way? Could professional historians learn something, beyond sheer readability, from the writers of fiction? People on both sides of the divide disagree.

Novelist Vanora Bennett rather resents the notion that historical fiction is an "unserious genre" focused on bodice-ripping and the sex life of royalty, when "in reality it's a very broad church that ranges from historical romances to War and Peace".

"I make every effort to ensure I have every possible known fact straight in my head before I start making things up," she says, "so as not to distort the framework I've chosen to work in. Because I tend to take a known period or story and look at it from an unusual angle - in my latest book (The People's Queen, 2010), for instance, the 14th-century credit crunch that happened in Geoffrey Chaucer's time, followed by the Peasants' Revolt, from Chaucer's point of view - I make a point of reading as widely as possible before and while writing."

Bennett's earlier novel, Queen of Silks (2009), is set in the time of Richard III, although the precise milieu was inspired by an academic article, "Two dozen and more silkwomen of fifteenth-century London", written by Anne F. Sutton, now historian emerita of the Mercers' Company. She would love to see "a less cautious exchange" between historians and historical novelists - "with perhaps more public discussion in which both species are invited to discuss a subject, bringing different kinds of expertise to the table".

Some academic historians seem willing to take up the offer.

Joanna Bourke, professor of history at Birkbeck, University of London, detests "the fluffy stuff of much popular history, wallowing in sentimentality, lauding either a nostalgic world of community, stability and certainty or, indeed, the opposite: all blood and guts, spilling over pages".

Yet she is also emphatic that "the best historical novelists teach historians a different kind of fidelity to the individual in history. These novels insist on the power of emotions over institutions, of contingency over the laws of causality."

ÁñÁ«ÊÓƵ

Helen Castor also believes historians have something to learn from novelists. A Fellow of Sidney Sussex College, Cambridge, she now devotes most of her time to producing well-researched books for a general audience - most recently She-Wolves: The Women Who Ruled England before Elizabeth (2010).

"I think that the presence of historical fiction can help keep us honest," she explains, "because, for all the care and rigour with which we need to treat our sources, we still have to join the dots to produce an interpretation of the past (and especially of human behaviour in the past) that we find convincing.

"The past is gone, and in attempting to summon it up in our own words, historians are engaged in a process of creative writing - creative writing according to different rules from writers of historical fiction, of course, but creative all the same. And I think that the explicitness of the process of creation in historical fiction can be a useful point of comparison and contrast with what we're doing; and also a reminder for historians that it won't do to treat our own protagonists as pieces on a chessboard rather than three-dimensional human beings living in a three-dimensional world."

It is probably unsurprising that historians can get proprietorial about the people and periods they study when novelists stray on to "their" territory.

"In general, I like historical novels," says Lucy Wooding, lecturer in early modern history at King's College London, "and I think they're a great way of enthusing people about the subject. But I find it hard to read novels set too close to home."

For example, she hasn't been able to tackle Hilary Mantel's Wolf Hall, which won the 2009 Man Booker Prize, "because I know it gives an entirely erroneous portrayal of Thomas Cromwell, and I won't be able to bear it".

ÁñÁ«ÊÓƵ

Yet her "real bugbears" are "dreadful 'historical' films, which play fast and loose with the historical evidence without even needing to do so. I can see how you have to simplify something in a screenplay, perhaps, but not gratuitously keep dead people alive and invent improbable sexual pairings when the history itself is already so fascinating.

"Let's face it, the Tudors hardly need to be sexed up or endowed with any extra dysfunction - the plain truth is gripping enough."

Ulrike Zitzlsperger, senior lecturer in German at the University of Exeter, agrees that "a good popular novel can achieve a lot in terms of raising awareness". Since fiction tends to keep returning to the same well-trodden times and places, she notes, it is not particularly difficult for authors to acquire some reasonably accurate local colour with the help of Wikipedia, dedicated home pages and a quick order of relevant books on Amazon. The question is what they do with it.

As soon as the setting is late-1920s Berlin, observes Zitzlsperger, "the average novel has its hero hasten over Potsdamer Platz getting hassled by the traffic (atmosphere); he then goes to Cafe Josty (location) where he meets/sees a famous actress (beginnings of film) who is later killed by a National Socialist (sense of doom) in one of the poorer districts (glamour versus poverty).

"Other predictable ingredients include the liberal use of cocaine (1920s decadence), anti-Semitism (a warning from history), advertising at night (modernity)...".

There is even a kind of retrospective fashion in hotels, with "the Adlon popular in that respect just now".

Even when novelists use reputable historical sources, adds Zitzlsperger, they tend to go to the same ones, often leading to "the reiteration of facts that foster the myths associated with certain periods - all these golden, roaring, swinging times".

No one can take out copyright, of course, on Tudor England or the Weimar Republic. But what happens when novelists draw more directly on the details of research carried out by individual scholars?

A girl lies prostrate, her lips touching the stone floor. A black cloth is thrown over her, and lighted candles are placed at her feet and at her head. Up above her, the litanies are being sung. All the signs suggest that she is dead. She is a witness at her own funeral."

This is the striking opening of Virgins of Venice: Enclosed Lives and Broken Vows in the Renaissance Convent by Mary Laven, who teaches history at Jesus College, Cambridge. Published by Penguin in 2002, it is written for a non-specialist audience, although with the full scholarly panoply of sources and footnotes, and uses court records to resurrect a lost world of intrigue, friendships and flirtations behind convent walls.

It soon attracted the attention of novelists. Both Michelle Lovric's The Remedy (2005) and Sarah Dunant's Sacred Hearts (2009), says Laven, "make use of Virgins of Venice for contextual detail - the sweets and biscuits produced by convent kitchens, the family visits to the nunnery parlours, the hairstyles fashioned by less committed nuns, the lap-dogs and chickens privately owned in contravention of the monastic vow of poverty. They also borrow more directly."

Lovric even starts her book with a clear echo of Laven's opening passage.

Although the direct borrowings are not huge and are duly acknowledged, Laven is clear that these novels could not have existed in their current form without her research and that of other scholars. She also spotted a few minor historical errors.

"Is there any reason", she asks, "why my colleagues and I should feel aggrieved by the use of our historical research in the cause of fiction? One objection may be that the novelists are profiting financially from our hard-won labours in the archives. Lovric and Dunant boast second homes in Venice and Florence, which is more than I do. But when academics experience frenzies of envy in relation to the high pay of other professionals (as they frequently do), novelists are rarely their number one target.

"A more probable source of discontent may be the misrepresentation of academic research. But if historical accuracy is our concern, then surely we should be grateful that novelists are making more rather than less use of our publications."

Laven adds that when she identified certain textual resonances between Virgins of Venice and the novels of Lovric and Dunant, "the echoes I heard were not of Laven, but of historical sources: nuns' writings, liturgical texts and trial records. If novelists choose to requote such titbits in order to lend authenticity to their characters' speech, then who am I to complain?"

ÁñÁ«ÊÓƵ