Thomas Dublin is a historian who interviewed coal-mining families after the mines closed. Among the people he'd interviewed were Tom and Ella Strohl. Some time later he was looking over photographs with them, and was surprised by what he saw. A great many of the pictures were of hunting trips, Tom out with his buddies. Dublin told him he hadn't realized how much hunting had mattered in his life. And Strohl answered, good-naturedly, "Well, you never asked."
As complaints go, it's hard to fault. The man had been interviewed by a professional, and a large piece of his life had got past the interviewer, because no question had gone looking for it. Linda Shopes, the historian who quotes the story, says that an interviewer's questions come out of what the interviewer thinks is important, and that more than one interviewer has had Dublin's experience.
The last lecture was about who is in the room, and what to keep out of a question. This is the opposite danger, the thing that never gets said because nothing invited it. So when it's you and your narrator, what do you actually say? And what do you do while she answers?
I'll go through it in the order an interview happens. The week before, and the list you bring. The question, which turns out to have very little in it. The silences, and the one method of interviewing I can show you a test of, which was built for the police. Then a quarrel among the professionals, over whether an interview is a conversation at all. And last, how to stop.
It begins before anybody presses record. Most of what follows comes from the manuals that archives, museums and universities write for their own interviewers.
More than a week ahead, talk to her without the recorder. It's known as a pre-interview, and its main job is to get the outline of her life: where and when she was born, where she's lived and what she's worked at, with the dates as near as she has them. Tell her that's all you're after, and that the stories are for the day itself. The Holocaust Memorial Museum's reason for the gap is that otherwise the recording fills up with reminders that she told you all this last week.
Then comes the homework. Say she grew up in a town you've never seen. Look it up, and find out what was going on there in her years. Draw a line down a page with her age set beside the events she lived through, because a child of eight saw a war differently from someone who was fifteen or twenty. The Shoah Foundation, which records the testimony of genocide survivors, asks its interviewers for at least half an hour of this, and as long again on the next step.
The next step is the list. By the list I mean one page of topics, a word or a phrase each, in the order of her life, and one line on yours will read, the house she grew up in. The manual from Baylor University wants keywords like that one and no written-out questions, because a question you've written down turns into an agenda.
On the day, begin with what's easy, her name and when and where she was born. The hard questions wait until later in the session.
One famous interviewer seems to say the opposite. Studs Terkel was the oral historian who cut the ribbon on the first StoryCorps booth, and when a reporter wanted to know how he did it, he said, "I never have a prepared list of questions." But a page of topics is not a list of questions. And when someone he was to interview on the radio had written a book, Terkel said, "I've got to read the book. And that takes time." For those guests, at least, he did the homework.
One rule sits over the whole page. When she goes somewhere the list doesn't, go with her. The list will keep until she's finished. A list can only hold what you already knew to ask, and by his own account Dublin hadn't known to ask Tom Strohl about hunting.
The middle of the session is made of questions, and a good one is shorter and plainer than you'd expect. The last lecture met the open question, which leaves the answer to her. The test for one is that a yes or a no won't answer it, and its use is that she steers you toward what matters most to her. Its opposite is a closed question, which a single word will answer.
Take that line on your list, the house she grew up in. The open question is this. Tell me about the house you grew up in. That's all of it. The easy way to spoil it is to stuff it, and ask where the house was, how many of you lived in it and whether it was cold. Ask several things at once and she'll probably answer only the first or the last, and the others are lost. So keep it open, keep it short, and ask one thing at a time.
Linda Shopes has a story from her own interviewing that shows what hangs on this. She was talking with an immigrant's daughter about her life in Baltimore in the middle of the last century, and she asked whether the woman had gone out to work after she married. The woman said she hadn't, and they moved on to her married life. Later, in passing, the woman mentioned that for several years of her marriage she had waited tables at a local restaurant in the dinner hour. Shopes asked about that. It turned out the woman had never thought of waitressing as work. She was, in her words, "helping Helen out". Helen owned the restaurant, and was a friend and a neighbor.
Shopes tells that story to show how a narrator and an interviewer can mean different things by one word. I'm borrowing it for another purpose. The first question carried the interviewer's own word for the thing, which was work. It got a no, and the no was honest. The woman brought the restaurant up herself. What the second question got was what the restaurant meant to her, and Shopes couldn't have written that question beforehand, because it was made out of something the woman had just said.
A question of that second kind is called a follow-up. It's made from something she has just said, and it asks for more of it. The plainest way is to hand her own words back. Go back to the house, and suppose that on her way to something else she mentions a lodger who had the front room. The follow-up is this. You said a lodger had the front room. Tell me more about that. And there you stop talking, because the question is finished. Another follow-up fits almost anywhere, and all it asks is what came next. And then what happened? You say that and no more. Follow-ups are where the depth comes from: she tells more, and she remembers further. If the lodger turns up in the middle of a story, don't stop the story for him. Write the word lodger in your notebook, and ask when she's finished.
Open and short is the rule, and it has three exceptions. The first is the closed question, which is mostly for checking a name or a date. It comes after the story, and not one upon another, or she'll feel she's being examined.
The second is the word why, on which the manuals split. The Shoah Foundation tells its interviewers to avoid it, because it can sound like a verdict on her. Yet the Baylor manual gives a why among its model questions, about why a family moved. I'd go by what the why is about, though neither manual sorts it that way. Ask why about what happened. About what she thinks, ask her to say more about her reasons.
The third is the long question, allowed in one place, where you change the subject. It comes in two sentences. The first says where you're going and why you're asking, and the second is the question. It sounds like this. We've talked about the house, and nobody in the family knows where you went to school. Tell me about your first school. That's as long as any question of yours should get.
That covers what you say: a page of topics you expect to leave, a short open question, and a second question you couldn't have written in advance. The other half is what you do while she answers, and for the most part the answer is nothing.
When she stops talking, don't speak. The Shoah Foundation's guidelines suggest counting to ten in your head. In their interviewers' experience she'll often add something, and what she adds is often the most revealing thing she says. Baylor's manual has a name for this, the silent question.
Ten seconds is longer than it sounds. Try it tonight against the second hand of a clock, and notice the number at which you want to speak.
It will feel like rudeness, ten seconds of nothing across the table from your own grandmother. But the silence belongs to her, because people need time to work out what they think. And it goes more easily if you look down and write a note.
Keep your own voice off the recording as well. In ordinary talk we encourage each other out loud, with little sounds of agreement. The microphone takes every one, and whoever plays the recording later will hear you on top of her. So do your encouraging in silence, with a nod or a smile. The manuals differ on how strict to be, and Baylor's allows a quiet, neutral word.
Then there's the moment she gets stuck, and the matter of how long you wait before you help. Wait out the count of ten first. After that it depends on what she has lost. If it's a word, the name for some object, and you know it, the Shoah Foundation's guidelines say to say it and ask her whether that's right. If it's a fact of her own story, a name, or what happened next, the last lecture's rule still holds, and you supply nothing. The guidelines' advice is to tell her it's all right not to have it, and then to try a question that might bring it back. Sometimes, they add, the kind thing is to ask whether she'd like to leave it for now.
The question to try is about what she saw and heard. It might be this one. If I'd been standing next to you, what would I be looking at? That asks for the room and leaves the fact alone. All of this is the manuals' experience, and they cite no tests for it. But a way of asking very like that last question has been put to the test.
In the 1980s police officers kept asking two university psychologists for a better way to question witnesses. What the two of them built they called the cognitive interview, and I'll call it the police method.
In one of their tests, people watched police training films of staged violent crimes. Two days afterwards each of them was interviewed by an experienced law-enforcement investigator. Some of the interviews were done the usual way. In others the investigator began by giving the witness instructions, and two of those matter to us.
The first went roughly like this. Before you tell me anything, rebuild the scene in your mind. Think about what the place looked like, where the furniture stood, what the weather and the light were like. Think about who was nearby, and how you were feeling. That was the first instruction. The second was to tell everything, including what seems too small to matter, because people hold things back when they aren't sure they're important. Later the two psychologists added rules for the interviewer: don't interrupt, let the witness set the pace, and tell them not to guess.
In 2010 a review pooled forty-six published papers on the method. Against an ordinary interview, it brought out many more correct details. It also brought slightly more details that were wrong, a coat called black when it was red. It did not bring more things that were never there at all. So a witness says a good deal more, about the same share of it is right, and there's more for you to check afterwards.
That result has limits. Most of it comes from laboratories and films. The figures are for the method as a whole: nothing in the review tests not interrupting by itself, and the count of ten isn't in it at all. And there's one thing tasks of this kind failed to do. Two lectures back, students' memories of the Challenger explosion had changed. What that lecture left out is that their interviewers also tried memory tasks from these psychologists' work, and not one student got the first version back.
A much-modified version of it has been tried at least once on old memories. People were asked about their daily physical activities thirty-five years before, and it got more answers and more exact ones, though that study's summary doesn't say whether anyone checked them.
Two things carry over to a kitchen table, and carrying them over is my doing. No manual mentions the police method, and I'm lifting two instructions out of a method that was tested whole. When the inventors tried their instructions one at a time, each one helped, and none did as well as all of them together.
Before a story that matters, take her back into the setting. Ask her to see the kitchen of the house she grew up in, where the table stood, what the light was like and who was in the room, and then ask about the day she left. That gives her a cue, a thing that sets a memory off, and it's the kind of cue that asks without telling. Tell her there was a stove in the corner, and you've supplied a stove. And don't ask her to imagine what must have been there. In one experiment, being asked to imagine a childhood event left people surer that it had happened. Ask only what she sees.
The other is to say once, at the start, that nothing is too small or too ordinary to be worth telling you. That instruction was written for witnesses to a crime, and it answers Tom Strohl before he has to say it. You're asking for everything, including what you didn't know to ask for.
Everything so far has kept you quiet, and some of the people who do this for a living think that's a mistake. StoryCorps, which writes for friends and families, tells you to think of it as a conversation, and to tell a story about her. Studs Terkel worked that way. "I'd say something about myself maybe, too, if it helps," he told an interviewer. A friend of his, a writer, said that anyone who knew Studs knew that "the guy just couldn't shut up".
There's a historian's version of the same point. Linda Shopes calls oral history, at its heart, a dialogue. Your questions draw out her answers and her answers shape your next question, so you are in the recording whether you talk about yourself or not.
On the other side is a list of tips by Willa Baum, one of the field's early leaders, in which the narrator is always he. Its first line says that an interview is not a dialogue, and the Minnesota Historical Society's tips open on the same line. You're there to get the narrator's story, and to say little beyond your questions.
One story belongs to both sides. Dave Isay, whose lost tape began this course, founded StoryCorps. Before that he gave tape recorders to two boys from the South Side of Chicago. One of them climbed into bed with his grandmother, recorder in hand, and Isay saw that the microphone gave the boy a license to ask her questions he had never asked before. The recorder and the occasion gave him leave to ask, and that's worth having whichever side you take.
I'll take a side here, and you should hear it as a choice. Two different things travel under the word conversation. One is what Shopes describes, that the two of you make the interview between you. I don't see how that can be avoided. The other is taking a turn yourself, with a story or a memory of your own, and that's the real disagreement.
If the recording is meant to be of her, my choice is her voice on it, and the conversation before and after. I have two reasons. The first is that the recording is being kept so that others can use it, and it's her they'll want. The second comes from the last lecture, which showed that a detail supplied by someone a person trusts is sometimes taken up as that person's own memory. Your stories, told in that room, are supplied details, and they can come back later as hers.
But if what you want to keep is the two of you talking, then StoryCorps has it right, and what settles it is something the first lecture had you decide before anything else, what the recording is for.
Whichever you choose, you don't have to be good at it. Terkel had a story for this. He's recording a woman in a housing project, say, and he forgets to press the button. She points it out, and he thanks her. In that moment, he said, he's no longer the man come down from Mount Olympus, and she feels herself his equal, or better. People accused him of doing it on purpose. "I'm just inept, simple as that," he said.
That leaves how to stop, and it has an order. Stop at about ninety minutes, and sooner if she tires: you told her about an hour. Before you stop, ask one last question. Is there anything I should have asked you about, and didn't? Then count to ten. Then thank her, and switch the recorder off. With it off, check the spelling of the names she used, and set a date for another session if she has more to say. And then, by all means, talk.
So, what do you actually say? Less than you'd expect: a question with nothing in it but the invitation, and then her own words handed back. And what do you do while she answers? You wait, to ten, without a sound, with a pencil in your hand. When a memory won't come, you take her back to the room it happened in.
No manual puts it this way, but I think those are one move, made over and over. The page of topics makes room for what you didn't know was there. So does the open question, and so does the follow-up. So does the silence, and so does asking for everything, however small.
Some of this isn't settled. The waiting rests on what interviewers have seen and the police method mostly on laboratories, and whether to call the session a conversation is a choice the profession hasn't made for you.
The easy questions went first, and the hard ones were left for later in the session. What you do when the story is painful, or when her memory is failing, is the next lecture.
Tom Strohl said it kindly, and it was true all the same. There's a question that goes looking for the hunting trips. It's the last one of the session, and it has no chance unless you then keep quiet long enough for her to answer.