I have been thinking about this a lot lately. I suppose one prompt is the latest series of The Tudors. I find it impossible to watch, not least because the actor playing Henry VIII looks nothing like Henry would have done at the age he had reached at the time depicted. It's just unreal, so totally divorced from my image of Henry that I can't relate to it.
Yet the series is popular, and I suppose it's interesting people in history, even if it's only the overdone era of Tudor history. (I can't image why this particular period seems so glorious to so many, but only World War Two and the nasty Nazis get more TV/movie coverage.) I don't really want to slag off The Tudors but talk about the issue of accuracy in general.
First, as I've said before, I think it's impossible to write a 100% accurate historical novel, and if by some miracle one could, the chances are that the result would be quite boring. Fiction is fiction. Even the best-researched novels are still novels, they tend to simplify events - let alone things like bureaucratic processes! - and in the last analysis they are written by people who are living in the present - post-Enlightenment thinkers. I can't really get inside the head of a medieval person. I can try, I can even explain the processes involved, talk about the social background, religious beliefs, upbringing, and so on. But at the end of the day, it's a best guess.
But granting all that, does it mean we shouldn't even try? For me, the first thing is not to knowingly change historical events. They are the scaffolding around which the story is told. I will not change the known outcomes of battles, jousts or parliaments, or make someone live significantly longer than they actually did. I'm extremely reluctant even to change a name, though it does get hard to differentiate between multiple blokes with the same moniker.
If you don't know, for a fact, where your character was on a particular day, then I think it's quite legitimate to move him or her someplace else. I've been trying quite hard to find where Richard of Gloucester went in 1469 after leaving Edward IV at Oundle, and before meeting up with him again in Yorkshire. I have some suspicions, but no proof. So I'm sending him to places that suit the story I am telling, one at least of which is inprobable, but not impossible. But it would be quite wrong to send him to Spain, for example, as he would not have time to get back. Equally, if he's known to be at Westminster on such and such a date, I don't really want to place him at Barnard Castle unless the story line makes it absolutely inescapable. Does it really matter in the context of a novel? Arguably it doesn't, but I prefer it that way.
Of course some historical events are disputed. Who kills the princes in the Tower, for example? One can gloss over this in a work of history. You can list three or four alternatives and come down on what you think is most probable. Or you can leave it entirely open. In a novel this is less straightforward. If you're writing from Elizabeth Woodville's point of view, for example, you may not have to give an outright answer. You can say what was reported to her and what she believed about it. It's highly unlikely that she knew what had happened. Richard III, on the other hand, almost certainly did know. A novel about him really does have to come up with an answer, one way or the other. (One about Elizabeth probably does as well, or the readers are going to be dreadfully disappointed. They will expect to 'know' possibly more than Elizabeth herself did!)
Background is also important. Medieval attitudes to marriage, for example, were quite different to ours. People of property did not expect to marry for love. Love was a bonus, a very desirable bonus admittedly, but definitely a bonus. Most people today would see Edward IV's marriage to Elizabeth as romantic - assuming he was not already married of course, and even then our modern attitudes might still forgive the power of love. His mother and Warwick almost certainly regarded the marriage as self-indulgent and irresponsible, and their attitude, not ours, would have been the commonplace one at the time. We may find their viewpoint unsympathetic, but we at least have to try to understand why they held it.
Then there's the little matter of religion. Religion was absolutely central to the lives of medieval people. We may be cynical, and suppose some of them at least didn't actually believe in it, but that's irrelevant. It still shaped their lives, impacted on the way they though, restricted what they could eat. Yet in so many novels of the middle ages this little matter is scarcely touched upon.
Fiction is fiction - but historical fiction needs to take into account the historical context. It would be absurd to write a contemporary novel about hill walkers in which everyone climbed the fells wearing carpet slippers and made up as clowns. Yet some historical novels are rather like that...