[RADIOLAB INTRO]

JAD ABUMRAD: Hey, I'm Jad Abumrad.

ROBERT KRULWICH: And I'm Robert Krulwich.

JAD: Yup. This is Radiolab. And today, just to start things off for this podcast, all right, let's just say that you love an author, but somehow the text isn't enough.

ROBERT: Okay.

JAD: Doesn't get you close enough to the author. So what do you do?

ROBERT: You know what you do?

JAD: What?

ROBERT: Take off your shoes.

JAD: [laughs]

ROBERT: You take off your socks.

JAD: As you're standing on the book?

ROBERT: And you stand on the book. And your whole body says, "Let me in!"

JAD: "Let me in!"

ROBERT: Yeah.

JAD: Well, let me give you a different flavor of that.

IAN LANCASHIRE: Hello.

JAD: How about you take the text, give it to this guy. He puts it into a computer, and you turn it into ...

IAN LANCASHIRE: Data. [laughs]

ROBERT: What do you mean? Who is that?

JAD: This is ...

IAN LANCASHIRE: My name is Ian Lancashire. I'm a professor of English at the University of Toronto.

JAD: Now Ian, as he said, is an English professor. But he's also a computer guy.

IAN LANCASHIRE: Right. Founded a computing center with the help of IBM Canada.

JAD: And the reason he combines those two is because he's interested in the secrets behind the author's words.

ROBERT: Hmm.

JAD: And that desire, he says, to take a text, spin it into data as a way to get into that author's head? Well, that goes back a long way.

IAN LANCASHIRE: Goes back to the fathers of the Christian Church.

JAD: To the Bible.

IAN LANCASHIRE: In the early Middle Ages.

JAD: Some monks decided to make what's called a "concordance" of the Bible. And what that means is they were gonna take every single word in the Bible—and there are 960,243 of them, at least in the King James version—and they were gonna list them all alphabetically, notate each time every single word was used and the context.

ROBERT: Ay yi yi.

IAN LANCASHIRE: Yes. Imagine it. You begin with the first verse.

JAD: In the beginning ...

IAN LANCASHIRE: [laughs] You create a heading for the first word.

JAD: "In."

IAN LANCASHIRE: And then for the second word.

JAD: "The."

IAN LANCASHIRE: And then for the third.

JAD: "Beginning!"

IAN LANCASHIRE: Every time you come across those words, you have to write a context.

JAD: In Genesis 1, verse 1, occurrence one, "In the beginning." Genesis 1, verse 6, occurrence two. "And God said let there be a firmament in the ..."

IAN LANCASHIRE: It's all handwritten. And at the end, you end—well, you end up with a lot of pieces of paper.

JAD: So many that it took those first monks who decided to do this an entire lifetime to complete it. Nowadays, you know, with computers, you can be done in ...

IAN LANCASHIRE: In under 15 seconds.

JAD: Bam! So that all just basically sets the stage for the story that I'm about to tell you. It's the 1980s. Ian is an English professor at Toronto. He's got a lab full of computers, and he's using them to analyze his favorite authors.

IAN LANCASHIRE: Samuel Taylor Coleridge, T.S. Eliot, James Joyce, Chaucer, Shakespeare.

JAD: And he's turning up some interesting stuff, sort of.

IAN LANCASHIRE: For example, in his poetry, Milton didn't use the word "Because." Who knows why?

JAD: Yeah. But at a certain point, Ian decided to look at more modern authors.

IAN LANCASHIRE: And so I turned to Agatha Christie.

JAD: At the time he was doing this—we're now in the '90s—Agatha Christie happened to be the most published author ever.

IAN LANCASHIRE: She sold a billion books.

JAD: A billion? Like, 'B' billion?

IAN LANCASHIRE: She was number one.

JAD: Behind God.

IAN LANCASHIRE: After the Bible, I think. [laughs] So what I did is I collected two of her earliest novels, written in the early '20s.

JAD: He fed those two into the computer.

IAN LANCASHIRE: Then I did the third.

JAD: Eventually, he would add in 14 additional books that cover 50 years of Agatha Christie's writing.

JAD: And what is the computer doing, exactly?

IAN LANCASHIRE: Measuring the individual concordance, word frequency, the vocabulary of the works.

JAD: And all the while it's spitting out these reports.

IAN LANCASHIRE: And I saw the totals at the bottom. So ...

JAD: Now first of all, the woman wrote 80 detective novels, which is just amazing in and of itself. The computer found that her use of language was relatively consistent and normal for the first 72 of those books. But something happened on book number 73. Something drastic.

ROBERT: What?

JAD: Suddenly her use of words like ...

IAN LANCASHIRE: Words like "thing," "anything," "something," "nothing."

JAD: What Ian calls indefinite words.

IAN LANCASHIRE: These words increased six times.

JAD: But also, when the computer added up the vocabulary size of that book?

IAN LANCASHIRE: That is how many different words are there in the first 50,000 words of a text?

JAD: It found in this book, there were 20 percent fewer different words.

IAN LANCASHIRE: That is astounding. That's one-fifth of her vocabulary lost.

JAD: It gradually dawned on him that what he might be seeing was the very beginning stages of an author losing herself.

IAN LANCASHIRE: She had developed Alzheimer's. I delayed publishing my results for two years. I had to have the results analyzed by a computational linguist and a statistician.

JAD: And in her lifetime, was she ever actually diagnosed?

IAN LANCASHIRE: Absolutely not. There was no diagnosis.

JAD: He said that some of her biographers suspected that something was up in her later years.

IAN LANCASHIRE: At one point, apparently, she cut off all her hair. She was not doing very well in interviews.

JAD: As far as we know, she was never taken to a doctor. Never got diagnosed.

IAN LANCASHIRE: I think her family closed around her and protected her. I realized that I was seeing something about the human mind. I was seeing the author in the text in a way that people hadn't seen the author in the text before.

JAD: Which raised a question for me, and I think this can apply to anyone. We all write a bazillion emails a day. I've got a decade's worth on my computer. Does that stuff hold clues about what will be, like early warning signs?

IAN LANCASHIRE: I think it's possible it does, yes. And it's well worth doing research about how a loss of vocabulary can be determined, let's say, in one's email over five or six years.

JAD: Indications are, he says, that those clues are there. Not only that, they may actually be there practically from the beginning.

IAN LANCASHIRE: A very famous example is the so-called nun study.

KELVIN LIM: Okay, the nun study actually began in 1990.

JAD: This is Dr. Kelvin Lim. He works at the University of Minnesota, and is the current director of the so-called nun study. And this study, more than any other that we know, really makes the point about the predictive power of the words we choose. The study began with a guy named David Snowden who wanted to look at aging over time. So he chose nuns because he wanted a group that was healthy.

KELVIN LIM: For example, they don't smoke, they don't drink.

JAD: They all have similar lifestyles.

KELVIN LIM: They obviously haven't had children.

JAD: So he approached this one particular order in Connecticut.

KELVIN LIM: Called "The School Sisters of Notre Dame."

JAD: And he signed up just short of 700 nuns. And the only stipulation being ...

KELVIN LIM: You had to be at least 75 years of age. And so we're now 20 years into the study, so that means the youngest of the sisters is about 95.

SISTER ALBERTA SHERIDAN: Yeah, I think I am. I am the youngest.

JAD: And you are 94 years old?

SISTER ALBERTA SHERIDAN: Yes, sir.

JAD: Not 95.

SISTER ALBERTA SHERIDAN: Not 95. [laughs]

JAD: This is Sister Alberta Sheridan.

SISTER ALBERTA SHERIDAN: I like the way you said that.

JAD: Do you happen to know who the oldest remaining sister in the study is?

SISTER ALBERTA SHERIDAN: Wait a minute. Now the one who was buried today, Jad, was 101. I think she was the oldest one in the study.

JAD: Wow!

SISTER ALBERTA SHERIDAN: In our province, yes.

JAD: The study began innocently enough, she says. The researchers would show up to the convent every year, give the nuns a bunch of tests.

SISTER ALBERTA SHERIDAN: Like, mostly for memory. Just questioning back and forth.

JAD: And then over the years as the nuns passed away—which many of them have at this point ...

SISTER ALBERTA SHERIDAN: They've all gone, Jad.

JAD: Of the original 678 sisters ...

KELVIN LIM: At this point. We have approximately 40 sisters still alive and participating in the study.

SISTER ALBERTA SHERIDAN: And I'm the only one left here in the Wilton Province.

JAD: And as the nuns would pass away, the researchers had arranged it so that they would get a small piece of their brains.

SISTER ALBERTA SHERIDAN: Yes.

JAD: Which they could examine for plaques and tangles.

SISTER ALBERTA SHERIDAN: Now this morning we buried a sister here, I told you. But the funeral was delayed a bit because she had to be taken to the hospital to have a portion of her brain removed to further the study.

JAD: Oh!

SISTER ALBERTA SHERIDAN: Mm-hmm.

JAD: Okay, so here's why I bring this study up: because of an accident that happened pretty early on that changed everything in the study. David Snowden, the main dude, was in the convent archives and he was talking to the archivist. The archivist says to him, "Hey, you know, all these nuns that you're studying who right now are over the age of 75? I actually have the essays that they wrote right when they got here."

KELVIN LIM: And they did this roughly at about age 18.

JAD: Like 60 years before?

SISTER ALBERTA SHERIDAN: Oh, yes. Oh, yeah.

KELVIN LIM: Right.

SISTER ALBERTA SHERIDAN: I have a copy of it at home. [laughs]

SISTER ALBERTA SHERIDAN: Oh, that's great. Come on in, Naomi.

NAOMI STAROBIN: Thank you so much.

JAD: We actually asked a reporter, Naomi Starobin to visit Sister Alberta at her home in Connecticut and have her read her essay that is now 76 years old.

NAOMI STAROBIN: Yeah, go ahead.

SISTER ALBERTA SHERIDAN: "Two days after the birth of the Christ child, I was brought as a belated Christmas gift to a Mr. and Mrs. Albert Joseph Sheridan of Providence, Rhode Island. A week later, the sparkly waters of baptism were poured over me." I'm not gonna read all this silly stuff that—when I first entered.

NAOMI STAROBIN: Why not?

SISTER ALBERTA SHERIDAN: It sounds kind of saccharin. I was only a teenager when I wrote.

JAD: But here's the thing: when the researchers found the essays like the one you just heard, it was a gold mine.

SERGUEI PAKHOMOV: It was a major, major find.

JAD: So they analyzed the essays looking primarily at ...

SERGUEI PAKHOMOV: ... two specific features of the language that was contained in these narratives.

JAD: That's Serguei Pakhomov. He does the analysis for the current nun study.

SERGUEI PAKHOMOV: In particular, they looked at the notion of grammatical complexity and idea density.

JAD: What is idea density? What does that mean?

SERGUEI PAKHOMOV: Idea density is a measure that looks at how many basic units of meaning are contained in any given utterance divided by the total number of words in that utterance.

JAD: In other words ...

SISTER ALBERTA SHERIDAN: The date of my birth is December 27.

JAD: ... like, if you were to listen to Sister Alberta's autobiography ...

SISTER ALBERTA SHERIDAN: When I was 11 years of age, my dear mother was called to God.

JAD: ... it's the number of little discrete ideas she's able to cram into one sentence.

SISTER ALBERTA SHERIDAN: This was to be a turning point in my life, as I had always had the ardent desire to become a sister.

JAD: Here's a classic example of the difference between low and high idea density. Here's low.

SERGUEI PAKHOMOV: From Sister Helen. "I was born in Eau Claire, Wisconsin on May 24, 1913, and was baptized in St. James Church."

JAD: Okay, that's low. Now here's high.

SERGUEI PAKHOMOV: From Sister Emma. "It was about a half hour before midnight between February 28 and 29 of the leap year 1912 when I began to live and to die as the third child of my mother whose maiden name is Hilda Hoffman and my father, Otto Schmidt."

JAD: I gotta say I'm liking the first one.

KELVIN LIM: Jad, probably you as a journalist is seeing the first one as straight to the point.

JAD: Yeah. It's good writing.

KELVIN LIM: And the second one seems kind of embellished.

JAD: A little bit, yeah.

JAD: But here's the punch line of all this: it turns out that the people who, when they were 18 wrote in that journalistically, very precise, low idea density sort of way, those people 60 years later were vastly, vastly more likely to develop dementia. In fact, based on those essays alone, the researchers could predict with about 85 percent accuracy what the nuns' brains would look like when they died and were able to look at the brains. I mean, would the brains have plaques and tangles that you associate with Alzheimer's or would they not?

ROBERT: [sighs]

JAD: What?

ROBERT: I mean, that's just crazy.

JAD: Wait. Why?

ROBERT: It's backwards reasoning. But we'll see, we'll see. I'm just—suddenly I'm suspicious. Here's a man who, from what you just said, has found the ones who got sick, and working backwards found certain incidences of this, that or the other and says, "Ah, this is a cause that produces this effect."

JAD: No, no, no, no. There's no cause and effect here.

KELVIN LIM: These studies are demonstrating associations, right? They're not demonstrating causality.

JAD: Right?

KELVIN LIM: It's a very important distinction.

JAD: This is just a correlation, okay?

ROBERT: But, you know, that may be one of 190 correlations that produce people who get Alzheimer's in the end.

JAD: Yeah. I mean, yeah. I mean, but let me argue your case actually from a different angle. Like, would this kind of linguistic analysis actually be relevant in the age of Twitter, where everything is short and clipped?

ROBERT: Well, people who Twitter don't only Twitter. They might also write small, short, dense essays for their ...

JAD: Yeah, but—well, you know, I guess you are right. It's like it's mostly about the thoughts in your head, not so much what you write.

ROBERT: Well, so what about Agatha Christie? Was there a conclusion about Agatha?

JAD: Yeah, there was.

IAN LANCASHIRE: Agatha Christie writing Elephants Can Remember.

JAD: This brings us back to Ian Lancashire and that 73rd book of Agatha Christie's that he analyzed and found that our vocabulary dipped. Well, before he did the analysis, he picked up that book and gave it a read. And like most people who read it, didn't like it.

IAN LANCASHIRE: Initially, I thought it was very poorly written, badly plotted, full of errors of time, of dating. Terrible read. Then I realized when I looked at the title, Elephants Can Remember ...

JAD: He realized that maybe Agatha Christie sensed what was happening to her.

IAN LANCASHIRE: ... she was responding to that truism that elephants never forget. The chief character is an aging female novelist named Ariadne, who is a foil for Agatha herself. And she—Ariadne—is suffering from memory loss.

JAD: In the story, she tries to help a detective solve this crime, but she has trouble because she keeps forgetting.

IAN LANCASHIRE: And the last sentence in that novel, in fact, is Agatha saying, "Well, maybe it's okay not to remember."

JAD: Wow.

IAN LANCASHIRE: She was trying to defend herself, defend her sense that she was forgetting, she was losing her vocabulary, she was losing her language. I began to see that Christie was heroic: still writing, despite this handicap.

JAD: Yeah.

IAN LANCASHIRE: And her willingness to do that at an age of 81, 82, struck me as heroic in a way.

ROBERT: Well, I understand that. The muse wouldn't quit but the tools all left the room.

JAD: Yeah. I think we should leave the room.

ROBERT: Okay.

[ANSWERING MACHINE: Message eight from phone number ...]

[LISTENER: Hello, this message is for Radiolab. My name is Paulina. I come from Chile. I'm a big fan of Radiolab, and this is your credits. The Radiolab podcast is funded in part by the National Science Foundation and the Sloan Foundation. Take care. Bye.]

[ANSWERING MACHINE: End of message.]

 

-30-

 

Copyright © 2023 New York Public Radio. All rights reserved. Visit our website terms of use at www.wnyc.org for further information.

 

New York Public Radio transcripts are created on a rush deadline, often by contractors. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of programming is the audio record.