Gödel Escher Bach, Douglas Hofstadter, and Me
In this essay I assume that you’re familiar with Douglas Hofstadter’s 1979 doorstop Gödel, Escher, Bach: an eternal golden braid (A metaphorical fugue on minds and machines in the spirit of Lewis Carroll). If you’re not, I expect you can find more interesting things to read than this.
A recent Medium essay by ‘philosophy geek’ Mark Johnson, Why Gödel, Escher, Bach is the most influential book in my life, caught my attention. That book, commonly referred to as ‘GEB,’ for which Douglas Hofstadter received a Pulitzer Prize in 1980, made a big impression on me too.
I remember the buzz that GEB generated when it came out — glowing reviews in Scientific American and the New York Times Book Review, among others — and I bought a copy as soon as it became available in Von’s Bookstore near the Purdue University campus. I read it with great enthusiasm. I found the ideas in it new and thrilling.
New and thrilling and way over my head. I probably understood about one third of that book my first time through it.
About a year later I was living in Somerville, MA, near the Tufts campus, with my fiancée, when I saw a notice that Hofstadter was to give a talk called “A Conversation with Einstein’s Brain.” So I went, with a friend of mine, and brought with me a copy of GEB for Hofstadter to sign. The talk, in which Hofstadter imagined Albert Einstein’s brain somehow encoded on a phonographic record, was predictably quirky, thought-provoking, and a bit irreverent. It was fun, as I recall.
During the reception after his talk I told Hofstadter that I intended to give a copy of his book as a wedding present to my wife, and asked him to write something apropos. Which he did.
Over the next twenty years I read GEB a couple more times, filling my copy (copies, more correctly) with annotations, placing post-its on seemingly every third page and putting notes on filing cards where they would’t fit on post-its or in margins. Like so many others, I became a GEB obsessive.
When I first met Doug Hofstadter at that Tufts lecture in 1980, I was only a few months into my job as a junior technical writer at computer maker Data General, a job I had talked my way into despite having no real qualifications. My understanding of the principles of computer hardware and software was, to put it mildly, rudimentary.
By the year 2002 I had written about twenty hardware and software manuals, had been a manager of a fifty-person information architecture group, half of which was in Silicon Valley, California, with the other half in Massachusetts, and I had written and published Acts of the Apostles, a nano-bio-cyberpunk novel about a Silicon Valley evil genius and would-be messiah and the cult of techbros who venerate him. Over twenty-two years I had acquired a much better foundation for engaging the ideas in Gödel, Escher Bach.
I still couldn’t stop thinking about that book. In particular I kept thinking about the authorship triangle that Hofstadter proposed on page 689. He imagines three novels such that the author of each one of them only exists in another novel in a series.
The caption reads, “There are three authors — Z, T, and E. Now it so happens that Z exists only in a novel by T. Likewise, T exists only in a novel by E. And strangely, E, too, only exists in a novel — by Z, of course. Now, is such an authorship triangle really possible?”
I decided that I would attempt to create just such a triangle. The second book in my series (which I call Mind over Matter (Overmind)) is a novella called Cheap Complex Devices (“CCD”), which purports to be the report of the inaugural Hofstadter Prize for Machine-Written Narrative. In CCD I imagined a novel written by a program like ChatGPT, the subject of which novel is a program like ChatGPT that runs amok as it writes a novel about machine-written narratives. (This was of course twenty years before LLM/transformative programs like ChatGPT existed, I was imagining what they might be like, and I must say that in many ways Cheap Complex Devices was right on the money.)
In writing Cheap Complex Devices I tried, as Hofstadter had done in writing GEB , to be “in the spirit of Lewis Carroll.” I put into my book, in one way or another, everything I knew or sensed or loved or misunderstood about the ideas of Douglass Hofstadter. I was — and still am, twenty years later — enormously pleased with what I had created. It was like nothing I had ever read before, and certainly like nothing I had ever written.
So I looked up Hofstadter’s address at Indiana University and mailed him a copy of Cheap Complex Devices, and I sent him an email telling him that I had done so.
A little while later I got a nice reply from him, in which he invited me to call him on the phone to chat. Which I of course did.
We talked about all manner of things; I don’t remember the details. One thing I do recall, clearly, however, is Hofstadter telling me that he was not going to read the little book I had written as a kind of homage to him. He didn’t have the time, he said. Life is short, and there were too many things he wanted yet to do.
I replied “It’s a very short book. You could have read it in the time you just spent talking to me!” But he had made up his mind and that was that.
On the basis of having written CCD, Salon asked me to write about the history and the then current state of “chatbots,” with a focus on The Loebner Prize, which was to be awarded to the first such bot that could pass a Turing Test. The result was an essay to which my editor gave the unfortunate title ‘Artificial Stupidity’: part one, part two.
During the course of my research for that story I had a few email conversations with Daniel Dennett, a ‘philosopher of consciousness’ based at Tufts who had been an early proponent of the Loebner Contest but later came to repudiate it. We talked not just about AI and chatbots, but about the history of the Loebner Contest and how it happened that the academic community that had once embraced Hugh Loebner and his prize later shunned him and denigrated his contest. I found some of Dennett’s arguments unconvincing and even a bit self-serving, and so in my story for Salon I kind of poked fun at him a little bit.
I need to mention here that Dennett was a friend an occasional co-author of Douglas Hofstadter.
Some little while after my Salon story came out got a note from Hofstadter inviting me to come to Tufts where he was to give a guest lecture in one of Dennett’s classes. After the class, Hofstadter wrote, he and Dennett and a few others would be going to dinner, and Hofstadter invited me to join them.
Remember, altough I had chatted with Dennett on the phone once or twice and had exchanged a few emails with him, I had never actually met him in person, and I had just ridiculed him (mildly, but still. . .) in an essay that had been read by tens of thousands of people. So I was a bit nervous.
But I did go, and it was a blast (Dennett was particularly gracious). Alas I don’t have time or space to write at length about it here. Highlights included: how, at dinner, Hofstadter pulled out a copy of the Salon story, which he had printed out, and read with great glee the portions where I had made fun of Dennett; how the next day, at MIT, thronged by a gaggle of admirers, Hofstadter spotted me and called out “This is John Sundman. He’s written a great article about AI on Salon. You should go read it;” and how when I put another copy of Cheap Complex Devices in Hofstadter’s hand, he said ‘thank you,’ but repeated his intention to not read it.
If you want to read all about how I researched and wrote that Salon article on Hugh Loebner and the state of artificial intelligence in the year 2003, or about that dinner with Hofstadter and Dennett et al, or about how I conceptualized my “Mind over Matter” trilogy as a Hofstadtertarian authorship triangle, I refer you to my substack, Sundman figures it out! In particular I refer you the three-part essay Scared Firefighter up in the Bucket, part one, part two, part three, where all these things, and more, are discussed.
The Wikipedia page for GEB includes these lines:
By exploring common themes in the lives and works of logician Kurt Gödel, artist M. C. Escher, and composer Johann Sebastian Bach, the book expounds concepts fundamental to mathematics, symmetry, and intelligence. Through short stories, illustrations, and analysis, the book discusses how systems can acquire meaningful context despite being made of “meaningless” elements [. . .]
Gödel, Escher, Bach takes the form of interweaving narratives.
Most essays in Sundman figures it out! begin with a paragraph that is a variation on this:
At Sundman figures it out! we aspire to the intricate art of consecution.
Themes emerge, interleave, dance around, fade away, and sometimes reappear. Incidents ramify, and their import may change upon being revisited. So reading earlier posts, for example this one, will enhance your experience of reading those that come later.
But if this is your first time with us, then, like William Shatner said, you have to start somewhere, so you might as well start here.
My substack, which I should probably subtitle “an autobiographical meditation in the spirit of Lewis Carroll and Douglas Hofstadter,” like GEB, takes the form of interweaving narratives. Interveavingness is of its essence.
Thus an account of my dinner with Doug Hofstadter and Dan Dennett appears in an essay about a scary experience I had when I was a firefighter — an essay that also touches on my wife’s porch gardens, the somewhat poignant tale of the last days of the Loebner Prize, a song by Jefferson Airplane, and the varieties of existential dread occasioned by recent developments in AI. I tell stories in the way they make sense to me. If I do it well, they make sense to my readers as well. Linear, they are not.
But I have to tell you, it makes figuring out how to come up with the best keywords to tag my substack essays with, and the right SEO summaries to give them so that search engines can find them, a right pain in the neck.
And for that, as for so much else, I blame, and I thank, Douglas Hofstadter.