Louise Doughty: The lady who ran off with the gypsies

Writing a novel about real people and actual events can feel like exploitation - especially when the characters are your own forbears. Louise Doughty describes her struggle to find a balance between invention and fidelity to the truth in her new novel, 'Stone Cradle'

The family travelled and worked as agricultural workers on the big Cambridgeshire estates, and it was when they were fruit-picking one summer that Tabletop met his future wife, a girl by the name of Hepzibah Lavinia Odell. Hepzibah was the landowner's daughter, "a high-born lady'', and her job was to go around collecting the rent from the gypsy caravans. She took one look at Tabletop and fell in love. When her father found out she was consorting with gypsies, he horse-whipped her and locked her in her room, but she climbed out of the window and shinnied down the drainpipe. Once her father got wind of her escape, he and her brothers went after her on horseback but it was too late. Hepzibah and Tabletop had eloped to Cambridge. None of Hepzibah's family attended the wedding and none of them ever spoke to her again but, without telling her, her father paid for the bells to be rung as they came out of the church.

Later, Hepzibah was to say, "quite often'', that it would have been better if one of the bells had fallen off and killed her stone dead, such a miserable life she had with Tabletop. They were on the road for a while, but so poor were they, they didn't even have a vardo (wagon). Hepzibah had been rejected by her own family but was never accepted by the gypsies either, on account of being a gorjer (non-gypsy). Her children were called half'n'halfs. It can't have been all bad between her and Tabletop, though, as they managed to have five children altogether, one of them my father's mother. But Hepzibah missed her old life and was always on at Tabletop to settle, and eventually the family moved into a terraced house in Old Gas Lane, a backstreet of East Cambridge. Tabletop's mother, Caroline Lee, also known as Spank, came to live with them for a bit. The two women did not get on. Later, the family moved to a village called Sutton, a few miles west of Ely, and later still to Peterborough.

When the Great War broke out, Tabletop joined the army and served in France, despite the fact that he was too old. His two sons joined up too, despite the fact that at least one of them was too young. All three survived, although one of the sons - my great uncle Cornelius - came back with such severe shellshock that he never recovered. When Tabletop was demobbed, he tried to live with Hepzibah again for a bit but it was obvious the marriage had broken down, so he went to live with his mother Spank.

A tough life came to a tough end for Hepzibah. She died in 1929, of stomach cancer, when she was only in her fifties. Her mother-in-law outlived her by two decades. When his mother died, Tabletop opened up his wife's grave and put his mother on top. I have visited the grave myself, in Eastfield cemetery in Peterborough. A loving wife / a mother dear / a faithful friend / lies resting here, reads the rhyme, and above, her name: Hepzibar Lavinia Odell. Misspelt.

At the bottom is inscribed, also of Caroline Lee. Spank may lie on top of Hepzibah, but she was clearly a bit of an afterthought as far as the stonemason was concerned.

I'm always getting ideas for novels by the sides of graves. (With my last book, Fires in the Dark, it was a mass grave in the Moravian highlands, in the Czech Republic, the burial site for several hundred Romanies who died in a nearby camp during the Second World War.) I don't know why this should be. Perhaps there is something about graves that fills a novelist's head with stories, a desire to tell the tales that lie beneath the earth. Or perhaps it is something rather less admirable: fear of death. Several novelist friends have owned up to me that they always work harder on a book when they know they are flying on an aeroplane soon, lest the plane crashes and their opus remains in its half-finished, parlous state forever.

As ideas go, basing a novel on one's family history has, at first sight, a pleasing simplicity: the ready-made cast of characters, the bare bones of a plot. But I soon discovered that there were other issues which make the process far more complex than I could have imagined.

Firstly, there is the basic question of names. If you use the real names along with real events then what, precisely, makes your novel a novel? The fact that you have made up bits you couldn't find out about, or elided over anything awkward? The beauty of your prose? Many a memoirist has done as much. But a memoirist is entitled - even obliged - to say in places, this much I am only guessing at. This much I could not know. A novelist has no such excuse.

If you don't change the names, do you have the necessary degree of freedom to change other facts? In my novel, the family are forced to flee the backstreets of East Cambridge after the main male character, whom I have called Lijah, gets in a fight with a local family. In real life, my Smith ancestors fled the city because of a smallpox epidemic, but smallpox didn't suit my plot.

Changing the names gave me permission to change other facts for creative reasons, to invent where I could not discover and to explain discoveries that did not suit my invention in whatever way I saw fit. One of my discoveries was that Hepzibah was certainly not a "high-born lady''. After many hours at the Family Record Centre in London, I found her birth certificate. I have no idea how she ended up in rural Cambridgeshire using the name Odell, but she was born Hepzibah Halfpenny, the illegitimate daughter of an illiterate laundry maid. This put me in something of a fix, as I had already written the section in which she elopes, based fairly closely on the family stories I had been told. It was one thing to invent based on a story, but could I invent based on a palpable untruth? So I then had to add chapters explaining how the illegitimate daughter of a laundry maid ends up collecting rent from gypsy caravans and is mistaken for the landowner's daughter. I could invent where I knew not, but not where I knew otherwise. It was an entirely illogical process, but it did give me the bulk of Part Two. Novelists are always doing this. We are fiction writers, therefore we act as if we have the Chance card marked, "get out of prison free". There is no leap or loop-the-loop we cannot justify.

Made-up characters. Made-up names. Instead of Caroline, George and Hepzibah, I have Clementina, Lijah and Rose. So I'm in the clear, aren't I? Well, not quite. There is another issue, quite separate from the creative ones.

A family history is never your history alone. It is collective. It belongs to a whole set of people, many of them still alive, and however respectful you try to be of other people's opinions and feelings, there is no getting away from the fact that when a history is written down, there is a sense in which that becomes the authorised version, simply because it is a version which can be read by strangers. A writer may protest until she is blue in the face that the Rose who dies of stomach cancer in 1929 in the novel Stone Cradle is not the same as the Hepzibah who died of stomach cancer in 1929 in real life - but my extended family may beg to differ. I got the idea from somewhere, and that somewhere is still attached to the idea.

The matter was further complicated in my case by the fact that, historically, Romanies and Travellers have been reluctant to share their family stories with outsiders. Many will use several names during their lifetime: an official name for dealing with the gorjers, the authorities, a nickname to be used by the wider family grouping, and a private name known only to a few. In very traditional European Roma society, such as the one I wrote about in Fires in the Dark, a child will be given a "real'' name, known only to its mother, which it is only considered mature enough to have possession of when it is older.

Spank and Tabletop are family names, private names, and as such I could not use them in my novel because it would have felt impossible, sacrilegious, to invent around those names. I thought long and hard about whether or not to use them in this article. I did it because I like the sound of them, and because I don't want them to die.

This, when it comes down to it, is my ultimate justification, and it is frighteningly close to that of a memoirist. No writer has ownership over their family history but it could be argued that a writer has responsibilities towards it that other family members do not. As a writer, you have a choice whether to let stories go, to remain in their graves, or to resurrect them and breathe life into them in your own particular fashion; flawed, invented, but recognisable.

What you don't have is any control over how that act of resurrection is perceived. Readers who don't know me will judge Stone Cradle purely on its strengths or weaknesses as a novel (thank God). My family will judge it within the terms of reference of what they know of the real-life people on whom it is based. They will decide whether it is an act of homage or a travesty, an insult or a gift. Most readers have to suspend disbelief, but they will have to suspend what they know. Perhaps, if I have done it well enough, they will. It is, after all, only a novel.

'Stone Cradle' is published by Simon & Schuster (£12.99) To buy a copy for £11.99 (free p&p), contact Independent Books Direct on 08700 798 897 or post your order to PO Box 60, Helston TR13 0TP