In Gustave Flaubert's great novel, Madame Bovary, the tragic heroine Emma is waiting at a grand house for a ball to begin when she is overwhelmed by a memory of childhood:
She saw the farm again, the muddy pond, her father in a blouse under the apple trees, and she saw herself again as formerly, skimming with her finger the cream off the milk-pans in the dairy.
There's nothing unusual about someone being surprised by a childhood memory; it would be an odd fictional character who wasn't sometimes ambushed by her own past. When you take a creative writing class, you are exhorted to bring your fictional creations alive by giving them believable internal worlds. As well as giving your characters motivations, secrets, fears, and ambitions, you need to give them memories.
This is something the great writers do instinctively. Hilary Mantel's acclaimed novel Wolf Hall is distinctive for giving vivid memories to its protagonist, Thomas Cromwell. A scene in which Cromwell has an erotic encouter in a sixteenth-century Cyprus gambling den is an event that the novelist herself could not literally have experienced. Again, no big surprise: writers invent, and they invent memories just like they invent other stuff.
It's impossible to say how much either writer based these imaginative reconstructions on their own experiences. It's likely that Flaubert will have had the experience of dipping a finger into a milk-pan, but he must also have rebuilt the memory from Emma's perspective, and made it true to her predicament. A novel is a device for enabling you to become characters you could never be, inhabiting places you could never possibly inhabit--for the author as well as the reader. And the creation of imaginary memories has, I think, something important to say about the science of memory.
I'm not here to tell you that reading fiction expands your brain; we already know that novels make the world a bigger, richer place, and it says something interesting about us when we seem to need neuroscience to confirm that fact. Instead, I want to persuade you that observing how an expert novelist constructs fictional memories tells us something about how memory works in all of us.
Some definitions first. We are in the territory of autobiographical memory, defined as our memory for the events of our own lives. The science of autobiographical memory is built on examining the errors we make when we recall events, and using that information to build up a picture of the complex machinery of remembering.
In the first stage of memory, encoding, we take in information through our perceptual systems and convert it into a form that can be laid down. The second stage, storage, preserves associations between those bits of information over days, months, and decades. Finally, at retrieval, we reconstruct the episode in question from its constituent parts. Crucially, we don't record events like a video camera for later playback; we reconstruct them at the moment we need to remember, from lots of different kinds of information stored at locations around the brain.
And here's where narrative comes in. Events unfold over time, and so do our memories of them. Narrative provides a temporal scaffold for putting the pieces of a memory back together. Some, in fact, have argued that acquiring the ability to do narrative allows young children to begin to tell stories about their pasts, explaining why, in adulthood, our very early years are lost to us.
To say that we narrativize memory doesn't mean that all our memories are fake, any more than it follows that a piece of reportage must be untruthful just because it has a story-like structure. It simply means that we tend to organize memories as narratives that involve characters with motivations, to which events happen and to which they respond according to their predicaments and personalities.
But that's not the only way in which memory is like fiction-writing. I've said that a rememberer constructs a representation of an event from several different kinds of information. That's also a pretty good description of how imagination works. Neuroscientists point to common systems in the brain underlying both imagination and memory, and an evolutionary argument is made for memory being as much about predicting the future as providing a detailed account of the past.
This is how a novelist can create a fictional memory with some elements that are real and others that are not. When we do the same thing with our own memories (as opposed to imaginings), those bits of information mostly correspond to how things really were--although the process of reconstruction means that unintentional errors can creep in. Psychologists sometimes ask their participants to make up memories for events that could have happened but didn't actually. Although it's quite difficult to distinguish naturally occurring false memories from genuine ones, experimental research shows that fabricated memories tend to be less vivid than true ones, and that they are created by modifying actual memories (as Flaubert presumably did with his fictional Emma).
I don't often put real memories into my novels, partly because I don't want to write about real people (they tend not to forgive that). I sometimes appropriate them from friends and family; a new strand of research shows that we actually do this quite a lot, especially with memories belonging to family members. Most often, though, I deal in made-up memories, for the simple reason that I haven't lived through the events I want my characters to have lived through.
Sometimes imagination and memory become confused. In Pieces of Light, I describe how I created a particular fictional scene set at an open-air swimming pool. When I went back to the city where I thought the pool actually was, I couldn't find it. I was confusing an imagining for a memory. Research tells us that simply imagining an event happening makes it more likely that you will subsequently (falsely) remember it. This effect is known as imagination inflation, and it can account for a variety of memory's tricks, from the persuasiveness of some advertising to the creation of false memories of abuse in certain forms of therapy.
We are all novelists when we are remembering. We start off with some sensory impressions of an event, along with factual knowledge about who we were then and what was going on for us. We put that information together in ways that reflect who we are now as well as who we were back then. Sometimes we put in information that shouldn't really be there: maybe it happened to someone else, or we heard about it on the TV, or we just imagined it. What results is a kind of story, a multisensory collage of different types of information. Flaubert was doing something similar when he gave Emma her memory of the milk-pans. In that sense, we are all Flauberts. Each time we cast back into the past, we demonstrate our capacity to narrativize ourselves.
Charles Fernyhough is the author of Pieces of Light [Harper Perennial, $15.99].