Historians produce imaginative fictions that convey truths about the past.
I think we all know this. We might not reflect on it very often, but most academic historians will have read Hayden White or Keith Jenkins. Like a stone in your shoe, you carry with you a sense of this, which is irritating. You might not think about their ideas when you get on with doing research and teaching from day to day, but, deep down, you would find it hard to say they are wrong.
But there are different kinds of imagination involved in academic research, and different disciplines have their own rules that set boundaries around what is allowable. These boundaries are constantly contested.
Take an example that is relatively easy to defend: the technique of ‘perhapsing’.
When a historian comes up against a crucial absence, a moment of great narrative or causal significance for which there is very little evidence, they deploy all their skills to research around this lacuna, and to provide a ‘perhaps’ that has to stand for what happened. This is the central technique in classic microhistories, such as The Cheese and the Worms, or more recent examples, such as Robert Darnton’s Poetry and the Police.
I read a particularly vivid example from the journalist Louise Kiernan. Kiernan was researching a story about a woman who was killed by a piece of falling glass. The piece she wrote included the following short paragraph:
No-one knows exactly how much time the glass took to fall – twenty-five seconds at most, perhaps as few as five. It may have floated flat as a table for a time or tumbled like a leaf, but gravity eventually pulled it into an angled or vertical position so it cut down like a knife.
Kiernan goes on:
To write that paragraph, I talked with two physics professors and two glass experts. There are calculations about gravity in that paragraph. While I was tempted to point out how hard I worked to get that information, I knew those sentences should stand on their own.
I suspect that an academic would have footnoted the research they did to produce this imaginative account – the standards of journalism and academia are quite different. But the idea that a fictional product of the writer’s imagination can depend so critically on research is common to both. This small fiction is the output of specialist investigation.
A harder example: within historiography is it ever acceptable to replace real examples with imagined ones?
The answer looks too obvious. Take the techniques of reconstruction often used by folklorists in the nineteenth century. When folklorists collected versions of stories such as ‘Cinderella’, or ‘Little Red Riding Hood’, they thought the oral performances they witnessed were imperfect recollections of older, grander, and purer narratives. They engaged in a process of ‘reconstructing’ the fullest versions of these stories, and inventing ‘originals’ that they thought were more accurate than the examples they had actually managed to collect from real individuals. They preferred an imagined tradition to the diversity and subtlety of the traditions they actually found.
This crosses a line, I think. It is a bait-and-switch with the evidence.
And yet, aren’t there times when it is acceptable to replace the specific examples we discover with composites, which take the details of many examples and produce fictional exempla?
At the ‘Creative Histories’ conference in July 2017 I was very struck by the way that the playwright Tracey Norman and the Circle of Spears production company did exactly this with early-modern witchcraft cases, producing a play that was based on one example, but drew on composite details of other cases. In a discussion after the performance, Tracey talked of how this allowed her to flesh out the story with realistic possibilities.
Which brings us back to perhapsing.
The point is not that historians prefer the ‘truthfulness’ of their own fictions to the specific truths that can be recovered from individual examples. It is that histiography produces both of these different kinds of truth. The truth of what happened in a specific situation is always mediated through a desire to turn those details into another story, an overall picture, a general theme, a social meaning.
And, for all their truthfulness and the richness of their meaning as a reflection of reality, the synthetic summaries that lie at the heart of many historiographical interventions are, nonetheless, imaginary.
 Louise Kiernan, in Mark Kramer and Wendy Call (eds.), Telling True Stories, p.147.