Douglas Hofstadter, Author of the "Legendary" GEB: Machines Could in Principle Have Human Minds — I Hope That Never Happens | Interviews

I don't like people calling GEB "the Bible of artificial intelligence." That's nonsense — it's simply insane.

"I don't like people calling it 'the Bible of artificial intelligence.' That's nonsense — it's simply insane. The book does argue that, in principle, the human mind could be simulated by a computer. I've always felt that way, and I still do. But I hope it never happens. The thought of that idea coming true doesn't excite me in the least. And the book isn't a Bible in any sense. It's a book about the brain, about philosophical questions."

That is Douglas R. Hofstadter, now 73, speaking to QDaily nearly 40 years after Gödel, Escher, Bach: An Eternal Golden Braid (GEB) was published.

Gödel, Escher, Bach came out in the United States in 1979 and became a legend. Across more than 800 pages, it braids together the mathematical logic of Kurt Gödel, the prints of M.C. Escher and the music of Johann Sebastian Bach, and out of that braid it draws the reader into mathematical logic, artificial intelligence, linguistics, genetics, music and painting.

The book caused a stir the moment it appeared, and the praise was extravagant. Scientific American's celebrated columnist Martin Gardner wrote: "Every few decades an unknown author brings out a book of such depth, clarity, range, wit, beauty and originality that it is recognized at once as a major literary event." John Holland, the first American to earn a doctorate in computer science — then still called "communication science" — recalled: "The people I knew were, for the most part, in a state of intense awe." Hofstadter, 35 at the time, won America's two most prestigious book prizes for this first book, the Pulitzer and the National Book Award, took a tenured chair at Indiana University, and for a while was the man of the hour.

The Atlantic wrote that Hofstadter seemed destined to become a permanent part of that cultural moment. Gödel, Escher, Bach was not merely a hugely influential book; it was a book entirely about the future. People called it the Bible of artificial intelligence. His descriptions of computer programs were accurate and inventive at once, and his account of "the secret software structure in our minds" sent a whole generation of young people off into artificial intelligence.

In the 1980s the book reached China as an adapted, abridged translation, published in the enormously influential "Toward the Future" series under the title GEB: An Eternal Golden Band, and it opened many young readers' eyes. Because it was an adaptation, though, that version carried only a fifth of the original. A complete Chinese translation did not appear until 1997, from the Commercial Press. It took more than ten years and nearly ten translators, with help from Hofstadter himself. The difficulty came mainly from his demand that the book be "transcreated": the translators were to be authors, rewriting the text with examples and language familiar to Chinese readers — and the book is packed with self-reference, palindromes, puns and acrostics. The Chinese subtitle, Ji Yi Bi Zhi Da Cheng, is one result: "Ji Yi Bi" is a phonetic rendering of "GEB," while "da cheng," a grand synthesis, is lifted from the Buddhist, philosophical and musical classics — the phrase echoes the original and works as a pun of its own.

Born in New York in 1945, Hofstadter is one of the world's best-known cognitive scientists. His father, Robert Hofstadter, won the Nobel Prize in Physics and taught physics at Stanford. Hofstadter grew up on the Stanford campus, "exposed to physics from a very young age" and "interested in physics." In 1965, at 20, he graduated from Stanford in mathematics. He went on to take a master's (1972) and a doctorate (1975) in physics at the University of Oregon, and then began his academic career in earnest.

Hofstadter has spent nearly all of his teaching and research life at Indiana University, where he is now Distinguished Professor of Cognitive Science and adjunct professor of comparative literature in the College of Arts and Sciences, and directs the Center for Research on Concepts and Cognition. Besides Gödel, Escher, Bach, he has written Surfaces and Essences: Analogy as the Fuel and Fire of Thinking, I Am a Strange Loop and Fluid Concepts and Creative Analogies: Computer Models of the Fundamental Mechanisms of Thought, among others. Fluid Concepts and Creative Analogies, published in 1995, happens to be the first book ever sold on Amazon.

For all its influence, Gödel, Escher, Bach landed at a turning point in the history of AI research. In the early 1980s the field was shrinking: money for long-term basic research kept drying up, and attention shifted toward systems that could be put to use. Hofstadter disliked all of it, casting "a skeptical eye on many widely celebrated AI projects and on the direction of the field as a whole," and found he had very little in common with most people working in it. What he thought about, always, was thinking itself — not utility, data and computation.

So he withdrew from what people called artificial intelligence. After roughly 1984 he largely stopped using the term for his own work and stopped attending AI conferences. He decided he was in fact a cognitive scientist, and that his research over the past 40-odd years had never left cognitive science.

"If I were to see genuine intelligence appear in a computer, I would be terrified. I am very afraid of that prospect, and I hope it never happens. I am not an advocate of artificial intelligence — I left the field 30 years ago — I am a researcher in cognitive science. For years I have studied how the human mind works, and, with the help of my graduate students, I use computer models to test my hypotheses. But I am not trying to build an intelligent computer. On the contrary, that goal is deeply opposed to what I do, and I will fight it with everything I have," Hofstadter wrote in a 2017 letter to the Chinese magazine 《大科技》(Da Keji).

For 40 years Hofstadter has been thinking about human thinking, and analogy is the key to his whole theory. His most recent book, Surfaces and Essences, appeared in 2013, written with the French psychologist Emmanuel Sander. In their account, cognition equals recognition, and recognition happens through analogy. So the capacity for analogy is the core of human cognition. Analogies are everywhere; we sail forever on a boundless sea of them. Without analogy there are no concepts, and without concepts there is no thought. People learn concepts through analogy, and think and talk by means of concepts.

After Surfaces and Essences came out, Liu Jian, taken with the book, went to Hofstadter and asked to translate it. Hofstadter arranged things so that Liu Jian and two Indiana University doctoral students became its Chinese translators. Liu Jian is now the librarian for the Department of Earth and Atmospheric Sciences at Indiana University. In the 1980s, as a graduate student in linguistics at Shandong University, he read Gödel, Escher, Bach in the original so avidly he "forgot to eat and sleep," and it "influenced me enormously." "How could he weave so many unrelated fields together so deftly! Of course there's plenty in it I don't understand. But reading it, I felt it was a work of astonishing brilliance among popular-science books. And he writes so well — reading it was deeply satisfying," Liu Jian told QDaily.

As with Gödel, Escher, Bach, Hofstadter wanted Liu Jian and his colleagues to transcreate Surfaces and Essences. So cheese in the book became tofu, red wine became tea, and so on. "He told us his translation philosophy at the start: you three are the authors. We were flattered, of course, but the pressure was enormous, and the work it took was absolutely huge. While we were in the United States we met about once a month to go through particular chapters — how should this sentence be understood? He would answer that for us. He also got David Moser to send us material. He helped us in every way he could," Liu Jian said. Moser is a close friend of Hofstadter's; many of the mistakes Moser made while learning Chinese turn up as examples in the Chinese edition of Surfaces and Essences. He also worked on the translation of Gödel, Escher, Bach.

Hofstadter also pays unusual attention to the form of language. When the English edition of Surfaces and Essences was published, Liu Jian said, he would not let the publisher touch the typesetting — it had to follow his layout exactly. Where each punctuation mark sat, how each paragraph was arranged, all of it mattered. Translating Pushkin's Eugene Onegin, he insisted on carrying over the verse form as well. For the final stanza he traveled to Pushkin's old apartment in St. Petersburg and sat there to translate it.

Hofstadter always carries a four-color ballpoint pen and a small notebook in his pocket. Since his teens he has collected his own slips of the tongue; the count now passes 10,000. The errors later became clues to how human thought works, and examples in his books. He is "extremely sensitive" to language, and will correct anyone's mistakes in earnest, almost never letting one pass. He has also studied Russian, French, Italian and Chinese, and translates on his own — Pushkin, Françoise Sagan and Yang Jiang among others.

"Forever in search of beauty."

That is how Hofstadter sums up his personal style and his goal on his own home page.

In search of what beauty?

"The beauty of music, of art, of words, of Chinese, of fiction, of sport, of photography, of physics, of mathematics… Beauty is everywhere," he told QDaily.

During the interview the white-haired Hofstadter kept breaking into a childlike excitement whenever he reached something he loved; he even turned down the lunch his publisher had arranged so he could talk with us a while longer. Whether the subject was The Catcher in the Rye, which he loves, or the salsa dancing he took up only after 60, he liked to describe it with the word "beauty," and to lose himself in it.

"He is a man who lives inside his own mind," Liu Jian said.

Q = QDaily

D = Douglas R. Hofstadter

Analogy Is the Core of Human Cognition

Q: Let's start with your new book, Surfaces and Essences, which is about to come out in China. I'm curious how it came about. Why did you write it?

D: Because it's a book about analogy. I think analogy-making is the core of the flow of consciousness and the core of thinking. My co-author, Emmanuel Sander, agrees. Analogy-making is not only a very complex, abstract, rare cognitive phenomenon — it is also extremely common. It runs through everything we think. We make many analogies every second: not once a week or once a day, but at the rate of one a second. In cognitive science this kind of thinking used to be treated as something very special, as if it were connected to royalty — high thought, thinking at the rank of a duke. (My name in Chinese is) Hou Daoren, and "hou" is also a word for a nobleman. Analogy was seen as a noble thought; the thinking of peasants, workers, ordinary people did not count. So Emmanuel and I both wanted the book to say as plainly as possible that analogy is in fact universal. We wrote two versions of it separately, one in English and one in French.

Q: Because your co-author is French?

D: Yes, and also because I love French — I have loved it since the first time I started learning it. And for two people whose views are very close, writing a book together was a wonderful chance to collaborate. The collaboration took far longer than we imagined. At first we thought a year or two; in the end it took eight years to finish. It was a long time, and hard work.

Q: From what I've read, the book's central claim can be summed up like this: the capacity for analogy is the core of human cognition; without analogy there are no concepts, and without concepts there is no thought; people learn concepts through analogy and think and communicate by means of concepts. You lay this out in detail in the book, but since most Chinese readers haven't been able to read it yet, could you explain briefly why you believe it?

D: Sure. In the book we work through some basic concepts. The first chapter uses the example of "mother," which fits the pattern of other concepts too. We start with the notion that there is one mother — only this mother, a capital-letter MAMA, and that every mommy in the world is that one. Then the child grows up, meets other children, and sees that those children also have someone similar beside them, playing a similar role. The word "mommy" stretches to cover "my friend's mommy," and becomes an analogy between one's own mommy and a friend's. Then, little by little, they realize that it isn't only children who have mommies — grown-ups have them too. Grown-ups have mommies! That is genuinely surprising. So even a mommy can have her own mommy. It's a delightful idea: mommy has a mommy! Children don't know any of this at first. They find it out later. Then they discover that animals have mothers too, and the concept keeps widening. Then we can talk about motherhood in an abstract sense — a company can have a parent company, for instance. In English we say "necessity is the mother of invention." We arrive at these abstract mothers. So the concept of "mother" goes from one very particular case to a very broad idea. It loses its uniqueness and becomes a concept that is everywhere.

Take another example: Chinese characters. Any character at all — say wo, the character for "I." It has one particular shape. The first time you meet it, it comes to you in a certain style of writing, a certain sequence of strokes. You copy those shapes, tracing them very carefully on paper. Chinese practice paper has grid squares, and you trace the character inside the grid. But after a while you realize that this shape is not so precise. It can take many shapes, and the variation grows more and more complicated, perhaps unpredictable. I have traveled around parts of China, to Hangzhou and Beijing, and photographed many signs with characters on them. There were plenty I couldn't read, so I would ask someone: what character is this? And they would say, that's wo. That's wo? What? How can that be wo? I have seen hundreds of versions of it, and I still couldn't recognize that one. But for a Chinese person, recognizing it is easy — obvious. I thought: how can that be? It's fascinating.

The same thing can happen with the letters of the alphabet. I could show you something and you would ask, what's that? I'd say, that's the letter F. And you'd think: how could that possibly be an F? Sorry, it is an F. The same thing is going on: recognizing every letter or every character is an act of analogy — an analogy with the shapes you have seen before. As with the mother example, you start from the simplest case.

Later it gets more complicated, extremely complicated, and the form becomes more flexible, more fluid. When you're thinking — for example, a little while ago Kelly handed me a sheet of paper. I can't find it now, but never mind. That sheet had some material for this interview on it. As I looked at it, in half a second I had read thirty characters. Half a second. I had to recognize every letter or character on it, which is like making 150 analogies in half a second. That is what I mean by analogy: it is everywhere. Analogy isn't only the abstract kind, "Einstein saw that this concept in physics resembles that one." Einstein spent months or years on a single analogy. That is one end of the range, the limiting case. It's like Mount Everest, the highest mountain in the world. The analogies we're talking about are only this high (gesturing very low). They aren't mountains; they're tiny. But they are the same as Everest, just in miniature. That is what we're trying to say.

Machine Translation Doesn't Understand Human Language at All — It Just Turns One Line of Text Into Another

Q: In January you wrote The Shallowness of Google Translate for The Atlantic, arguing that Google Translate is a failed project, full of errors, with no understanding of human language at all — and that a bigger database won't fix it, because what it lacks is a human mind. Why do you say that? And what is a human mind? Could you spell out your view?

D: What I mean is that Google Translate is a product, the result of thousands of people working hard to build an enormous database. It uses bilingual databases plus neural networks, and it goes around anything at the level of understanding — it avoids letting the computer truly know anything. They are trying to make the computer translate, not understand. In many cases Google Translate does a decent job, and in some it is impressive. But in other cases it fails, and falls into absurdity, over and over.

This isn't only Google Translate. It's the same with Baidu's and Tencent's translation; there's a German company called DeepL, and Microsoft has one too, and they all fail, in many cases badly. Because there is no understanding, and no attempt at understanding. They try to route around understanding, to avoid it. Fine — in some cases their performance really is impressive, and very fast. And you can type something into the box, or rather you don't even type it, you copy some text and paste it in. Two seconds later out comes a result. What you put in comes back to you in another language. We call the translated text "it," but "it" isn't the same thing as the original. After machine translation there is certainly something there, but it may not correspond to the original at all. It might be entirely wrong, it might be nonsense. Yet it claims to be the translation. Sometimes it translates very well, impressively so. I'm not saying it is all bad. What I'm saying is that this is deeply unreliable, and that it has nothing to do with real translation work.

Q: You say it lacks a mind. What does having a mind mean?

D: It means understanding concepts. Take Google Translate: it doesn't understand the world at all. It doesn't know there are tables in the world, doesn't know human beings exist. It has no idea what's out there. It doesn't know there is a universe. It doesn't know now and then, here and there. It doesn't know what motion is. It knows nothing. All it has is a pile of English text and a pile of Chinese text, and all it knows is that this stretch of text corresponds to that one. That's all. Even "know" is too generous, because it has no capacity to know — but it has an algorithm that can turn one line of text into another line of text, and that is the whole of it. No knowledge, nothing. It does its job very efficiently. It looks as though it understands something, but that isn't what's happening.

Q: You also wrote in that piece: "In principle, machines could one day think, be creative, and even feel nostalgia, excitement, fear, elation or resignation, or hope — meaning that one day they could gracefully carry out the task of translating between languages." Why do you say that? Could you say more?

D: In my first book, Gödel, Escher, Bach, I had to say very carefully that I believe thought — human thought, animal thought, any thought — is governed by the laws of nature. In my experience, as a physicist, the laws of nature are mathematical laws. That means they can be implemented in a computer program; you can build a computer program that does anything at all. We haven't done it. But I don't want us to get there. The thought of that idea coming true doesn't excite me in the least.

In principle, though, it is absolutely possible. What the brain does is a computation. In theory, I mean, the brain is itself a computer, and that is what it does. So any physical entity could have a mind of its own, whether it is a computer or a brain. That is entirely possible. But to have a mind it must interact with the world — it has to. It can't be like a database translating fragments of text; that is not interacting with the world at all, it has nothing to do with it.

Q: If what you describe came true, and machines had minds, emotions and experiences like ours, do you think there would still be a difference between people and machines?

D: No difference.

Q: And how do you feel about that outcome?

D: Let me leave your question aside for a moment and say something else. When I started out in the 1970s, thinking I would become an AI researcher — that's 40 years ago now — my view was that computers were obviously stupider than people. They had no mind at all. That's how they were. They played chess worse than human players; they were only better at crunching numbers. In many ways they performed terribly. And my feeling then was that even if it didn't take thousands of years, it would probably take a few hundred before computers came anywhere near human intellectual thought.

But I said in the book that in theory this is entirely conceivable. About 20 years later, in the 1990s, some developments worried me: progress might be running faster than I had expected. That was exactly what I did not want to see. IBM's Deep Blue beat Garry Kasparov at chess — he was the world champion in 1997. That made me very sad. When the Korean player Lee Sedol was beaten by AlphaGo, I felt worse. I was very sad and low at the time. I am not someone who wants to see this happen. I don't approve of it, but I think it is possible.

It is entirely possible, but it requires interacting with the world and having feelings of one's own. To become like a person, each computer would also need an individual identity. It couldn't be one of a run of identical products — you know how many products are indistinguishable from one another. If you like, it would need its own life, its own experiences, its own personal style. It would have to be an individual. That probably won't happen soon, but it could happen. I hope it doesn't. And it is conceivable that the process takes much less time than I expected. I don't like the idea.

Q: In the past few years it seems more and more people are talking about an "AI threat."

D: I'm not sure they know what they're talking about.

Q: They argue that AI as it is developing will affect privacy, the job market, weapons of mass destruction, and could even destroy humanity. What do you make of that?

D: Well, there's no way for us to know what will happen. On the one hand, I don't want to see computers become smarter than people. On the other hand, you could look at it this way: this human world is full of aggression, hostility and hatred, and people kill each other. There is a great deal of hatred in the world. Suppose that computer — though I don't like the word "computer" — suppose some artificial beings were created that were kinder, more altruistic and more grateful than we are. Maybe that would be better. If they were kind, so much the better: get rid of these stupid humans who are forever killing and hating each other, and you'd have a kind community in the world, getting along.

Q: What if some of those artificial beings turned out to be Nazis?

D: Right, that could happen too. Who knows? I don't know. I think it would be terrible. How would you control it? None of us could.

My Style and My Goal Are to Search Forever for Beauty

Q: An article in The Atlantic mentioned that since about the age of 15 you have reread The Catcher in the Rye every ten years. Could you say why you love the book so much, and what it has done to you?

D: I can't explain why I love it. But perhaps I can give you some broader impressions. It's the story of a boy in turmoil, an extremely honest person who sees phoniness everywhere. "Phoniness" means fake, artificial, dishonest. He feels the world is full of dishonesty and hypocrisy. It disappoints him deeply, because phony things are everywhere.

But he also sees innocence. In Central Park he watches small children riding a carousel. You know the thing, with the wooden horses, and the children taking turns, going up and down. The carousel turns, up and down, and he sees innocence in it. He sees innocence in his little sister too. He believes strongly in the power of innocence, so he tries to rescue and protect innocent things.

That is what the title The Catcher in the Rye is about. Rye is a grain; in Chinese I think you'd say a great field of wheat, which is close enough. The idea behind the title is that a lot of children are playing in this field of rye, and beside it there is a cliff. The protagonist, Holden, stands at the edge and catches the children before they fall. The cliff stands for phoniness and the field for innocence, for purity, for what hasn't been spoiled — and he catches the playing children before they go over. That scene moves me deeply, but I can't explain what makes it so beautiful.

Q: Apart from The Catcher in the Rye, looking back over your life, which other books have mattered most to you? Why?

D: That's hard to say. Many books have influenced me in different ways, and I've been asked this many times. The first that comes to mind is a physics book: Waves, by F.S. Crawford (Frank S. Crawford), volume three of the Berkeley Physics Course. I love it because it is beautifully written, explaining tremendously great ideas in physics in a very down-to-earth way. It is one of the most beautiful books I have ever read.

Q: Your page on Indiana University's website says: "Hofstadter sums up his personal style and personal goal as follows: forever in search of beauty." Why put it that way? What kind of beauty?

D: The beauty of music, of art, of words, of Chinese, of fiction, of sport, of photography. I'm very interested in Chinese art. This is my second time in China; the first time I didn't really notice this, but this time I noticed Chinese trees. Some of them are very different from the trees I've seen in America. Their shapes are different. The way the branches of some trees move in the wind looks to me very like the shapes of Chinese characters. It's hard to explain, I'd have to give examples. There's a kind of tree — I don't really know its name. I'm not sure whether it's called a dragon-claw or a dragon-horn tree, and there are many of them along West Lake in Hangzhou. That tree is wild, quite extraordinary.

Q: You see beauty in that tree's branches?

D: Yes, and I think I see the same beauty in Chinese calligraphy, and in English calligraphy too. And the beauty of music is everywhere. All of it is beautiful, anyway. Beauty is everywhere. The beauty of physics, the beauty of mathematics… I should have mentioned physics and mathematics earlier; I don't know why I forgot them at first.

I'm Translating Yang Jiang's We Three, and I Want to Call It Woman Sad

Q: Speaking of Chinese — I hear that besides Russian, French and Italian you've studied Chinese too. What has learning it been like?

D: Well, when I was five, six, seven years old, still very small, I lived with my parents near San Francisco. They took me into the city now and then. There's a very large Chinatown there, and we used to go. I saw Chinese characters there and heard some Cantonese on the street, though not Mandarin. I was fascinated. I loved those characters. I loved how mysterious and strange they were, most of them square, and they left a very deep impression on me.

So, long ago, in my thirties, I started studying Chinese. And I studied hard, because it is an extremely difficult language, and I did all I could, within my limits, to master it. Let me use Mount Everest again. Chinese is the Everest of languages. I'm still climbing very low down, but the mountain itself is very tall.

Q: You've done a lot of literary translation — Pushkin, Sagan and others — written a book about translation, Le Ton beau de Marot: In Praise of the Music of Language, and you care a great deal about how your own work is translated. What are your principles of translation?

D: There is no universal principle. Every text you translate is different. Every book is different, and different books call for different principles. You can't boil it down to one rule that covers everything.

Take We Three, which I'm translating now. Start with the title. How should it be said in English? We could call it We Three, or Three of Us. But I have a very different rendering. I think We Three is a decent translation, of course, but I don't want to use it. Other people may tell me I should, because it's the obvious choice. I want to do it differently: I want to call it Woman Sad. "Woman" stands for wo men, "we," and "Sad" is sa, "three" — and this Woman is Yang Jiang, who is very sad in the book. Have you read it? You know how the story opens with that terrifying dream: Qian Zhongshu is taken away somewhere, and Yang Jiang goes to see him, back and forth along that long road. Their daughter falls ill and goes into hospital, and later the daughter dies, and Zhongshu dies too. It is truly sad. The whole story carries that sorrow.

So my title is Woman Sad, which I find rather fun. But I'm not sure about it either. Maybe my friend David Moser thinks it's a bad translation. My wife doesn't like it. I like it. I think it's an interesting way to do it. We'll find out eventually. My point is that when you translate, when you think about what the original means, a rendering like Woman Sad can catch something of the flavor of the Chinese title. Woman Sad, wo men sa. I just find it interesting.

What do you want to catch? What do you want to keep? When you translate, what are you trying to do? The words for mom and dad, for instance, are used constantly in the book. Do I want to put "mommy" and "daddy" in their place? Or do I want "mama" and "baba"? Right now I think I'll use mama and baba, because "daddy" isn't a Chinese notion. "Daddy" belongs entirely to English-language culture. It isn't the right word for the image of a Chinese father.

Q: Then how do you get American readers to understand what mama and baba mean?

D: Mama is obvious — we say mama too. As for baba, you only have to explain it once. The first time the word appears you put its meaning in parentheses. I never use footnotes. It's obvious, I'd say: the book mentions a daughter, and then says two people have a daughter, so of course those two are the mother and the father. But there are many other cases in the book. Sanlihe, for instance — do I say "Three Mile River," or "San Li River," or keep the pinyin, "Sanli He"? Or gu yi dao, the old post road — do I say "Old Post Road," or "Gu Yi Road"? These are all interesting questions. How much of the Chinese do I actually want to keep?

Q: What do you plan to do?

D: I don't know yet. So you see there is no principle as such. You still have to think it through carefully, get used to the process, try the options — what you think while translating, and what other people think of your choices.

Q: Do you ask your wife and friends?

D: Yes, I ask them. So I really have no fixed principles of translation. I have to listen to other people, follow my own instincts, and work out how many elements rooted in Chinese culture I want to keep.

Q: While still making sure American readers can follow it.

D: Exactly. All English-language readers, in fact — American, British, Australian, Canadian, any country, anyone who reads English, including Chinese people who read English.

I Don't Like People Calling GEB "the Bible of Artificial Intelligence." That's Nonsense

Q: Gödel, Escher, Bach is still your most influential book in China. Looking back at a work you finished nearly 40 years ago, what strikes you most about it?

D: It's a book full of life. It brims with vitality, and at the same time it is serious. It's also a book that sets people thinking about the mind. One of the key things about it is that the chapters alternate with dialogues. The dialogues are great fun. What's distinctive about them is that the characters in a dialogue often talk about the very structure of the dialogue they're in.

The reader notices this, of course, but the characters don't. That's the fun of it — it's like a game, played in a way that's very abstract but also funny. I wrote the book in a state of excitement. I think it passes that excitement on to readers. I don't like people calling it "the Bible of artificial intelligence." That's nonsense — it's simply insane. Though I understand why they'd think so, because the book gets to artificial intelligence fairly early on. It does argue that, in principle, the human mind could be simulated by a computer. I've always felt that way, and I still do.

But it isn't a Bible in any sense either. It's a book about thinking, the mind and consciousness. What is the "I"? The self, as Chinese puts it. It isn't a book about computers; it's a book about the brain, about philosophical questions. It's been more than 40 years since I wrote it — and I don't mean the publication date, it came out in 1979, but it was written in the mid-1970s, so it's more than 40 years old.

Q: Your father was a scientist too. What effect did he have on you?

D: Well, I was exposed to physics from a very young age, because my father was a physicist. He had many colleagues and students who came to the house. Even as a small child I loved physics, simply because so much of it was around me — not because my father wanted me to be interested in it, but I became interested anyway. Later I realized he was a passionate physicist who loved physics, loved the beauty of physical law, and passed that passion on to me without either of us noticing. That's one way to put it. But my mother had a large influence too — her love of words, for instance, her love of wordplay.

Q: Was she a linguist?

D: No, she just loved language. My parents met through music; they both loved jazz. Because of them I loved jazz from an early age too, and classical music. Their influence on me was about equal, though my mother's leaned toward the humanities and my father's toward science. Both of them shaped me deeply.

Q: What have you been thinking about lately? Anything to do with your next book?

D: I can't tell you what my next book will be.

Q: Then what have you been thinking about?

D: I'm always thinking about a lot of things.

Q: Questions about the mind?

D: Thinking about thinking. It has always been that question, really.

Q: What has changed most in your life in recent years?

D: I learned to dance salsa, which was great fun. Unfortunately my wife and I don't dance any more. We stopped a few years ago. But we did dance — five years of salsa. The feeling was wonderful. That's another subject about beauty: the beauty of dance, incredibly beautiful. The movement of the body, and the way your body synchronizes and coordinates with another's. Not just one other person, but possibly a whole group of people dancing with you. A group in perfect coordination. Ah, and then, at a certain moment, beauty just happens. Beautiful and strange.

Q: Has your sense of beauty stayed the same over all these years?

D: Well, I think I've always loved dance, but I didn't learn to dance until I was 60. So yes, I've always been this way — you could say my sense of beauty is the same — but I only took part in dancing myself after 60. Before that I always believed dance was beautiful; I just hadn't danced.

Q: I was asking about your sense of beauty in general, not about dance in particular.

D: Oh. Then it's much the same.

Q: I've noticed you like wearing clothes with butterflies on them. Why? Does it have anything to do with "Hofstadter's butterfly"?

D: I don't think so. I bought my first butterfly shirt probably some thirty-odd years before the phrase "Hofstadter's butterfly" existed. I made a discovery back then, but I never called it anyone's butterfly; I always called it the G Plot. So I never thought of it as a butterfly at all. To me the theory was just the G Plot. It was many years after that first butterfly shirt that the phrase "Hofstadter's butterfly" caught on. I do like butterflies, of course. Because in a sense they are innocent. Like Holden Caulfield in The Catcher in the Rye. I like innocent things.

Banner image: Hofstadter, from Indiana University; cover image from pxhere.