Critics are always provoked by the use of facts and 'real' people in novels, so it's time for rules to be agreed, writes John Lanchester
'The imagination is in love with the feel of fact," John Bayley once wrote. This is a subtle point, and cunningly worded: the feel of fact is not necessarily the same thing as fact tout court; something can feel like a fact without being one, and a good few things which pretend to be facts don't feel exactly true. Bayley's second memoir, Iris and the Friends, caused a mini-scandal when it turned out that he had fictionalised and stylised some of the characters. There was an irony here, since other people had complained that his first volume of memoirs, Iris, was too unsparing in its account of Iris Murdoch's life and decline into Alzheimer's. His first memoir got people going because it had too much fact, his second because it had too much fiction.
The area where fact and fiction overlap is murky and contested. The questions would be much simpler if fiction could simply, regally, ignore the world of fact, as some critics occasionally call for it to do. Fiction should be lofty, interior, superior; commerce with the world of brand names, street names, dates and places is beneath it. This is a version of the novel which sets it apart from contemporary reality, and makes it a separate form of artistic discourse that has more in common with itself than with what is happening in the world. In other words, a novel should be more about its own relationship with Dickens and Thackeray and Turgenev and Bellow than with what will be on the news tomorrow morning.
The favourite tag of this movement - half-slogan, half-mantra - is T.S. Eliot's observation about Henry James that he "had a mind so fine no idea could violate it".
This is an honourable and internally coherent view of the novel. The first problem with it is that it makes the novel sound boring, and self-absorbed. The term "literary fiction" encapsulates this argument in a way, by implying that there is a kind of fiction which is specifically literary and more interested in its own literariness than in anything about the wider world. For the wary general reader, the term has come to translate as: "Warning: Prize Entry; Do Not Read."
The second problem with the Eliot-James view is that it doesn't correspond to the reality of the great novels that have actually been written - which tend to be more various and more interesting than accounts of the novel make them sound. I would argue that the six greatest novels ever written are Moby Dick, Madame Bovary, Middlemarch, War and Peace, Ulysses, and À La Recherche du Temps Perdu. Of these books, Proust's is the only one not consumingly interested in questions of fact. Madame Bovary was based on the marriage of Eugene and Delphine Delamare, and on her suicide.
Moby Dick is based on the story of the whaleship Essex, sunk in mid-Pacific by an unprecedentedly aggressive sperm whale, and the book has an astonishing - a grotesque, obsessive, hilarious - amount of factual information about whales.
Middlemarch imagines an entire fictional town, from its sewerage systems to its municipal elections; George Eliot's research involved extensive reading in town planning, geography and everything else. Tolstoy immersed himself in the history of the Napoleonic wars to write War and Peace, which did not prevent the book's factual inaccuracies from being a point of controversy in its many terrible reviews ("one continually got the impression that a narrow-minded but garrulous corporal was boasting of his exploits in some remote hamlet to a group of gawping hicks", wrote one critic).
And no work of fiction is more preoccupied with fact than Ulysses, which is, perhaps not coincidentally, in many people's view the best novel ever written. One of Joyce's ambitions for Ulysses was that if Dublin were destroyed it could be rebuilt from scratch from the novel, so extensively specific was the factual information about the city.
The reason why so many great novels have been so intricately interwoven with fact is straightforward: because fact is interesting. Melville was fascinated by whales, Flaubert by the Delamares, George Eliot by the phenomenon of the modern city, and so on. The task of writing novels, it seems to me, is to be as interested in the world as in the form. One of the bad things that has happened to novels in the last few decades is that an interest in the world has come increasingly to be taken as the prerogative of commercial fiction. Literary books are about themselves; commercial books are about the world around them. When a literary book turns out to be a big commercial success, it is almost always a book with a large dollop of external, worldly, factual or quasi-factual reality; two recent examples are Zadie Smith's White Teeth and Jonathan Franzen's The Corrections, both of which are full of information. A novel is supposed to be novel, to be news - otherwise it would be called an "oldvel".
But given this centrality of fact to the history of the novel, it is strange that the rules about how to use fact in fiction are so vague and improvised. Nothing is more likely to provoke a row than the use of fact in a novel: for example, it was asked whether Thomas Keneally's Booker-winning Schindler's Ark was a novel at all, whether Robert Stone borrowed too freely from Nick Tomalin's book about sailor Donald Crowhurst for his Outerbridge Reach, and whether Andrew O'Hagan's Personality drew too closely on the life of Lena Zavaroni. A shared characteristic of these rows is that they aren't triggered by readers, whose sense of the interplay between fact and fiction is more robust than that of critics.
But the fact-and-fiction question is easy to latch on to, and latch on to it critics do - usually, as in John Bayley's case, with the complaint that there is either too much or too little fact or fiction. When was the last time you read a piece that said a book used factual and fictional elements, and got the mix just right? The absence of rules means writers have to make things up as they go along, and resolve these questions by feel. There are models of how to do it, and how not to. Don DeLillo's novel, Libra, draws heavily on the Warren Report, and is a masterpiece; Oliver Stone's film, JFK, tried to tell the secret history of the Kennedy murder, and is a pernicious, mendacious disaster. John le Carré's Tinker, Tailor trilogy fictionalised the story of Kim Philby and his colleagues, and made up a compelling new myth about modern Englishness; the recent BBC series, Cambridge Spies, used the real names but fudged the real stories, and the result was a howling dog.
From brooding on these questions, I have come up with one simple rule. If a character in a novel borrows the name of a real person, the external attributes of the person's life have to be adhered to as closely as possible. If your book has someone in it called Clement Attlee, it has to be Clement Attlee. If it isn't, you are merely trying to borrow the glamour of fact, and are certain to end up with something fudged and confused. The same with place names, dates and the rest of the furniture of the real.
But if your novel draws on real people in order to depart from them - as most of the great novels mentioned above do, or as Orson Welles did in Citizen Kane - then you signal the change by using a new name. Fact is interesting, but fiction is interesting too. A new name means you are free to make up anything you like; reality is a point of departure, not a destination, and you have no further responsibility to history, only to the reader. To anyone who finds this rule too robust, the reply is simple: tough. If you want non-fiction, go and read some.
- Guardian Service
John Lanchester's novel, Fragrant Harbour, is published by Faber