Why did Charles Dodgson (Lewis Carroll) fall out with the Liddell family?
How did the cliff dwellers of Montezuma Castle manage their days?
You need to answer these questions, but no one knows the answers. What do you do? You make informed guesses based on the information you have been able to discover. And there’s nothing wrong with doing that, because this is fiction.
I asked Nancy Jardine, whom I interviewed here and who writes – amongst other genres – historical novels set in Celtic/Roman Britain, for one example of when she had to make an informed guess. This is her response:
I’ve had to do a fair bit of that in my Celtic Fervour Series, the reasons being that the only source materials for the times are Greek or Roman – and biased at that. The Celts left no written evidence at all – save what’s found on stone inscriptions. In Book 3 of the series, I have a large battle at a place named Beinn Na Ciche (Gaelic for a hill range that’s nine miles from my home). Bennachie, as it’s currently called on the map, is only one possible site earmarked by historians for a battle between Romans and Celts that the Roman historian Tacitus wrote about. The thing is that Tacitus may only have been attempting to make his father-in-law Agricola seem like a greater general than he really was. None of the other Scottish contender sites – for what was later named the Battle of Mons Graupius by early Victorians – have provided conclusive evidence. I made my ‘informed’ choice and decided that the topographical information given by Tacitus matched the landscape around Bennachie, but even more important for me was that in 2006 local archaeologists decided that the Romans had had some 30,000 soldiers in Durno – a marching camp opposite the fooothills of Bennachie. The number of soldiers (which is drawn from archaeological evidence and not written evidence) was sufficient for me to use the site in my novel. In fact most of what I’ve written as historical facts on Celtic life is drawn from purely archaeological sources and is therefore all interpretative.