A few things I've noticed about writing characters who lived in centuries I didn't.
Min Jin Lee spent nearly 30 years working on Pachinko. Sunja, the protagonist, doesn't speak much in the novel. Her silence is period-accurate: women in early 20th-century Korea didn't have many words available to them. The characterization comes through what she does. She pickles cabbage, she carries loads, she makes decisions by acting on them instead of announcing them. You learn who she is by watching her hands.
The costume gets the first draft. The person gets the second. Most early attempts at historical fiction characters read like someone wearing a period outfit at a Renaissance fair, saying all the right words in all the right patterns, while the actual human underneath stays vague.
Edward Rutherfurd writes novels that span centuries through the same families. London and Sarum follow bloodlines across dozens of generations. The trick is that each generation's protagonist has to feel like a distinct person, not a reincarnation of their ancestor. Genetics gives you the continuity. Circumstance gives you the difference. A blacksmith's grandson who becomes a merchant in a different century shouldn't walk through the world the same way, even if they share a stubborn streak.
Historical characters don't know they're historical. They think they're contemporary. This sounds obvious but it changes everything about how you write their inner lives.
Philippa Gregory wrote The Other Boleyn Girl and made Mary Boleyn, the footnote, the protagonist. The character readers remember from Tudor history isn't the one Gregory chose to center. That inversion is worth thinking about: the most interesting character in any historical period might be the one the history books barely mention. The kings and generals already have their stories told. The scullery maid, the second wife, the translator nobody credited, those people had inner lives too, and fiction is sometimes the only form that can give it back to them.
Your character's relationship to hygiene, to food, to weather, to the passage of time, to pain, to boredom, to darkness after sunset, all of it was different, and the differences aren't just scenic details you can sprinkle in, they're the architecture of a completely different way of being in a body, and if you don't let that reality settle into the prose then you've basically written a modern person in an old hat.
I'm still not sure whether it's better to let historical characters use slightly modern speech patterns or to commit fully to period-accurate dialogue. Full accuracy can feel alienating on the page. But modern phrasing can break the spell entirely. I've seen both work. I've seen both fail. I don't think there's a rule here, just a series of judgment calls that you have to keep making, sentence by sentence.
Viet Thanh Nguyen's narrator in The Sympathizer is a spy with two minds. He's half-French and half-Vietnamese, working for the Communist cause while embedded with South Vietnamese refugees in America. The dual consciousness becomes a way to write a character who sees both sides of the Vietnam War without the novel having to pick one. That split identity lets Nguyen do something almost impossible: write a protagonist who is genuinely sympathetic to contradictory worldviews, and who suffers for holding both at once.
Every historical character carries the weight of what the writer knows will happen next. You know the plague is coming. You know the revolution will fail. They don't. Writing well in that gap, between your knowledge and their ignorance, is one of the hardest things about the form.
Min Jin Lee once said, "I write about people who have been erased." That single sentence tells you more about character selection in historical fiction than most entire craft books. The deeper question for any historical fiction writer is whose story got lost, and whether fiction can give it back.