In one bound volume at the Library of Congress there are two typed interviews, a few pages apart, with the same woman. Her name was Susan Hamlin. She lived at 17 Henrietta Street in Charleston, South Carolina, she had been a slave, and when the interviewers came, in the 1930s, she was past a hundred by her own count.
Nothing was recorded, so what I'll be quoting is each interviewer's writing-up of what she said. In one of them, her master is a bank president named Fuller, and she says, "Mr. Fuller was a good man and his wife's people been grand people, all good to their slaves."
In the other, her father is whipped. A washerwoman in the household is beaten until Hamlin thinks she'll die. Mothers and children are sold apart, and people, she says, were always dying of a broken heart. And there's one flat sentence: "I had some terribly bad experiences."
Even her age differs, and her name. She's Susan Hamlin, a hundred and four, in one, and Susan Hamilton, a hundred and one, in the other.
It's the same woman, at the same address, talking about the same Mr. Fuller. It isn't the same visitor. At the head of each typescript is the name of the person who came to ask.
The last lecture left this behind: everything an old person remembers has been told before, to somebody, and the listener has a share in the telling. So how much does the person asking change the story? And what follows from the fact that the person asking is going to be you, one of the family?
The route is this. Who the asker is, which the typescripts show. What the asker says, which psychologists have measured. And then your own case, where the person asking is the narrator's grandchild. Each of these leaves you something to do. I'll give it as it comes up, and at the end in the order you'd do it.
Start with the one where the master is a good man, because its author wrote down how she got it.
She was Jessie Butler, a white woman collecting for the Federal Writers' Project, whose interviewers wrote up life histories from what people told them. One July day she came to Henrietta Street asking for Susan Hamlin. Hamlin was sitting just inside her front door, on the step up to the porch. When she heard a caller asking for her, Butler writes, she assumed it was someone from the Welfare office, which had given her aid until it closed. Butler's next words are these: "I did not correct this impression."
They talk. Hamlin says that when the Civil War began she was hired out to mind children, at seven dollars a month, and that her master got the money. Butler asks her, "Don't you think that was fair?" And she presses it: if he fed and clothed you, shouldn't he be paid? Hamlin answers, "Course it been fair."
Then Hamlin volunteers something. About a month ago, she says, a man came who said he was from the Government, sent to find out about slavery. He asked all kinds of questions: didn't the white people do this, did they do that. And Hamlin says, "I give him most a book, and what he give me? A dime." Most of a book, that is, for ten cents. But her master was a good man, she goes on. She tells Butler, "I ain't going to let nobody lie on him."
The other typescript is by Augustus Ladson, who was Black, and Hamlin calls him son. Butler's ends with a note that the name is Hamlin and not Hamilton, which she'd had spelled out for her. Hamilton is the name on Ladson's, so I take his visit to be the earlier one. And I take him to be the man with the dime, who is never named.
What Ladson took down is another slavery. Take the washerwoman, Clory, who didn't take foolishness from anybody. One day the mistress came into the laundry and found fault with the clothes, and Clory picked her up and threw her out of the door. The mistress was pregnant, and the baby was born within two hours. Clory begged to be sold, but this was a master who would never sell his slaves, and what she got was a whipping. Hamlin doesn't say who gave it, only that she thought Clory would die. Clory got well, and meaner, and in the end the master hired her out.
Her father is there too. Another man owned him, and whipped him one Sunday. On the Monday he gave that master the same, and got away north on a ship.
Set the two side by side. The bank president is in both, and so are the seven dollars. In both, this master sells none of his own people, and in both, other people's children are sold away from their mothers. To Butler she says the mothers would scream and holler, and that you know how they carry on. To Ladson she says that people were always dying of a broken heart. It's no clean experiment: these are two different pens, and Butler let a mistake stand and put her own answer into her question. But most of the difference is what Hamlin chose to tell, and what she'd say it meant. The seven dollars is called fair only in Butler's, and Butler's is the one that asks.
It wasn't only her. One study went through the collection, and some earlier interviews, and counted. Of the former slaves interviewed by whites, 72 percent said the food had been good. Of those interviewed by Black interviewers, 46 percent did. These were different people in different places, and not one woman asked twice. But I can't believe that the ones who drew a white interviewer had been that much better fed.
Not everything moved with the interviewer. A second study looked at how family relationships were reported, and found no sign that the interviewer's race affected it. What moved was what lay between a former slave and a white caller: how she'd been fed, and what she'd say of her master. I'll call those the charged subjects.
So you can't take Butler's typescript as it stands, and you can't throw the collection out as spoiled. Linda Shopes, the historian we've met in both lectures, has a better question, and she puts it to every interview. Who is talking? Who are they talking to? What for, and in what circumstances? A narrator sizes up whoever is asking, and works out what's safe to tell them, what they ought to be told, and what to leave out.
It's a question for any account you already have, a tape in a drawer or a story your aunt tells. Who was it told to? Hamlin thought she was telling the Welfare.
So much for who Butler was. Now what she said. In the last lecture a question was a cue, and a cue decides which part of a life answers. A question can do more than that. Go back to Butler's: don't you think that was fair? You'd never ask anything as crude as that. Probably not.
In an experiment published in 1974, a psychologist named Elizabeth Loftus showed students short films of car accidents. After each film came a questionnaire, and in it Loftus put the one question that mattered: "About how fast were the cars going when they hit each other?" Other students got the same question with a different verb, and there were five verbs in all. Where the question said hit, the average answer was thirty-four miles an hour. Where it said smashed into, it was about forty.
That is the famous half of the experiment, and it has been hard to repeat: a study of 1,204 people in 2024 found no difference. So don't rest much on the speeds. Loftus didn't stop there.
Say you're a student in her next experiment. You watch a film that runs under a minute. You're asked the speed question, with the word smashed. A week later you come back, and there's no film this time. You're given ten questions, and one of them is, "Did you see any broken glass?"
What do you say? There was no broken glass in the film. Of fifty students who'd been asked with the word smashed, sixteen said yes. Of fifty asked with the word hit, seven did.
Nobody was guessing at a speed this time. They were saying what they had seen. The question had supplied one thing, that the cars smashed, and the glass followed. It doesn't take a crude question. In that experiment, a verb did it.
One experiment is thin ground, as the speeds have shown. The wider finding isn't thin. In hundreds of experiments since, by Loftus's count, something supplied afterwards has turned up in what people said they saw.
I'll call what the question put there a supplied detail: something that comes from somebody else, afterwards, and is kept as if it had been there all along. Loftus's account is that what was seen and what was supplied join up in memory, until you can't say which piece came from where. Others hold that the memory isn't touched, and that the question only shapes the answers of people who never took the detail in. Either way it's on your recording.
Lawyers have one name for Butler's question and for Loftus's. A leading question is one that tells the person what answer is wanted, by its form or by what's in it. Butler's leads by its form. Loftus's leads by what's in it, the word smashed.
So keep your own answer out of the question. The giveaway is a tag: the words wasn't it on the end, or the words don't you think on the front. Take the tag off. Better still, ask an open question. An open question leaves the answer to her. Hamlin had just said that her master got the money. The open question was four words long, and Butler never said them: tell me about that.
None of this bans photographs. A cue asks, and a supplied detail tells. Put a photograph on the table and ask who this is, and that's a cue. Tell her who it is, and you've supplied the answer.
Can a whole event be supplied? In a later study of Loftus's, twenty-four people read stories from their own childhoods, collected from close relatives, and one story was invented: at about five, they'd been lost in a shopping mall. Seven of the twenty-four then remembered being lost, in part or in full. The method leaned on the family's word that it had happened. How often it works is still argued over. One pooled count of such studies says three people in ten. Two critics, reviewing them in 2016, put it at about half that for a memory the person is sure of. By either count it can be done, and to a minority.
In the film and in the mall, most people did not take up what was supplied. Even so, I'd do a sum that the experiments didn't. A supplied detail takes best when time has let the memory fade, and somewhat better, in most studies though not in all, in the old than in the young. An old memory, an old narrator and a relative's word add up to a family interview, and in that room the relative's word is yours.
So supply nothing. When your narrator hunts for a name or a year, or stops short of an ending you've heard a hundred times, don't hand it to her. How long to wait before you help is the next lecture's business. And get her account first, on the recording, before anybody else's is mentioned: yours, her sister's, the one the family tells. If another account has to be put to her, the manual from Baylor University has you tell her that some people say it happened otherwise, and ask what she thinks. I'd keep even that until hers is safe.
That is the asker in the answer. Who they are changes what a narrator chooses to tell, and what she'll say it meant. What they say can put something there that wasn't there before. But every asker so far has been a stranger, and what moved for a stranger in the 1930s were the charged subjects: what a former slave would say about a master to a white caller. Between you and her, the charged subject is the family.
For this half I can't offer you a comparison, the same grandmother with a stranger one week and a grandchild the next. It rests on a ranger's experience, a historian's judgment and one son's article. Begin with a relative who was only sitting in.
Manzanar, in California, was one of the camps where the United States held Japanese Americans during the Second World War. It's a historic site now, and Rosemary Masters was the ranger running its oral history program when she set this story down. The people she interviewed were in their eighties and nineties. Some of them, she says, had never spoken about it to their own families, and often she learned, usually afterwards, that she was the first person to hear it.
Her practice was to spend a long time with a narrator before any recording. Once she didn't, because a nephew wanted to make the arrangements himself. The nephew also sat in. What the ranger didn't know was that uncle and nephew didn't talk openly with each other. They reached the part of the story where hard things had happened to the family, and the nephew had never heard any of it. The uncle asked not to go on. The ranger calls it the nephew fiasco. It was the only interview she had ever had to stop because of what was being said and who was there to hear it, and when she wrote, the uncle still hadn't finished his.
All the ranger says the nephew did at the interview was sit in. With no talk beforehand, that was enough, and the lesson she took from it is your instruction. Arrange it with the narrator yourself, and not through the aunt she lives with. When you arrange it, say it'll be the two of you. The Baylor manual makes that a rule, announced when the interview is set up, which gives you grounds to ask everyone else to leave when the recorder goes on.
Not every guide insists on this. Densho, the oral history archive, tells families to consider group interviews. I'd sort it as Shopes does, who has often thought that relatives interviewed together mostly put the family on record. A session with the family gets you the family, and a session alone gets you the person. And a table of relatives is a table of people supplying details. So alone first, and if someone has to stay, Baylor's ground rules are that each question goes to one person, and nobody answers for anybody else.
Now move the relative into the interviewer's chair, which is where you'll be sitting.
Shopes's own example of a narrator sizing up an interviewer is your case exactly: a grandparent talking to a grandchild for a family history. The less savory parts may well get held back, she writes, to protect the child, to set an example, and to keep up what she calls the family's myths. I'll call what comes out the official version.
One relative has written down what it's like from the interviewer's side. Maarman Tshehla was a scholar at the University of South Africa, and in 2015 he took down the life of his mother, who was eighty-two, for a book for the family. He was the eighth of her ten surviving children. Afterwards he wrote an article looking back on how he'd done it.
Being her son got him something. He knew there was a question to ask. He'd once seen a birth certificate, for a brother a year younger than himself who was never spoken of before that day or since. When he'd asked, he was told he was too young to deserve knowing. By 2015 he'd been her son for more than four decades, and he was old enough. The article doesn't say what she told him.
Being her son also cost him, and for each cost there's something you can do.
She took it for granted that her listener knew the places and the cousins, and he had to fill them in himself afterwards. When your narrator skips what you both know, ask anyway, on behalf of whoever hears this later. You might say this to her. I knew Grandpa. Your great-grandchildren never will. Tell me about him as if I'd never met him. That's all you'd say. An oral history is kept so that others can use it, and a question like that one is how the others get in.
She gave him the official version, and he couldn't press. She wanted her life to do good to those who read it, so whatever didn't teach was left out, or smoothed. And as a much younger son there were limits to how hard he could push, which he puts down to his own culture's rules. So afterwards, let somebody else have a turn: an old friend of hers, a cousin from the other side, someone her own age. Charleston ended up with two interviews with Susan Hamlin, and you can set out to get two.
And he couldn't stop being her son. His article ends on this sentence: "If only I were more qualified and more eloquent and not so close to her." There's no cure for that. There is one thing to do. At the start of the recording, say who you are to her, and who else is in the room. Densho asks families for exactly that: the people involved, passive listeners included.
The model I'd take for it is Jessie Butler, of all people. She got her interview by a deception, and a later lecture has what you owe a narrator before you begin. But the one thing Butler did right was to write the deception down. That sentence of hers, that she did not correct the impression, is why anyone can read her typescript for what it is. You can't take yourself out of the recording, so put yourself into it. As the recorder starts, you say something like this. This is her granddaughter asking, and it's just the two of us. That will do, and a later lecture has the rest.
Both typescripts are Susan Hamlin. How much does the person asking change the story? What she chooses to tell, and what she'll say it meant, change a great deal. The facts change less, unless the asker supplies them. The first of those is a choice she makes about you. The second is something you'd be doing to her memory, and you can decline to do it.
A stranger wouldn't know there's a birth certificate to ask about. You would, so I wouldn't wait for a better interviewer.
What follows from being family is a way of going about it. Beforehand, fix it with her directly, for the two of you. When the recorder goes on, say who's asking. While she talks, ask so that the answer is hers, hand her nothing, and ask on behalf of the stranger who'll listen one day. And afterwards, let someone else have a turn.
That has been about who is in the room, and what to keep out of a question. What goes into one, and what you do while she's answering, is the next lecture.