15 Whiplashed by the Manichean Struggle Between the Two Cultures
I began writing Machines Who Think in the mid-1970s even as I was interviewing the pioneer scientists, and my agent distributed a book proposal. No one in New York publishing knew what artificial intelligence was. Once the topic was defined for them, they failed to see why it was important. “We’ve already published a book on computers,” one editor replied to my proposal in 1974. “Too bad it’s too late,” another said cryptically, as he too rejected the idea. You can’t really blame them. The whole idea was preposterous—Machines? Thinking?
These Delphic utterances discouraged my agent and deeply unnerved me. Was it all my imagination? Andrei Yershov at the Doctor program terminal, opening his heart to a nonjudgmental computer interrogator? The Logic Theorist inventing a better proof to a theorem than the brilliancies of Alfred North Whitehead and Bertrand Russell? The robot arms swinging autonomously within their protective plexiglass cages at Stanford and MIT? A waist-high Alberich, shake, rattle, and rolling around the halls of the Stanford Research Institute? It was called Shakey, as if we’d returned to some long-ago, less tactful age when humans named each other bluntly—Fatso, Shorty, Gimpy, Cruikshanks. Shakey wobbled its way around, evading walls and people, veering to a plug when juice ran low.
Machines all around me at Carnegie Mellon, MIT, Stanford, and Stanford Research Institute were busily performing all sorts of tasks: playing chess, playing poker (though not as well as they would by 2017, when two different programs beat professional human champions at Texas Hold ‘em).
At Carnegie Mellon, the Cheese Co-Op program took monthly orders for cheeses unavailable at the Giant Eagle, the local supermarket. The program then calculated how much cheese of various kinds must be purchased at the wholesale cheese market in the Allegheny River’s Strip District. The program did a further calculation of how to cut up wheels and wedges optimally to meet each family’s order. The project’s popularity soon meant hundreds of pounds of cheeses were delivered to the campus. (Joe said silent prayers that the funding agencies didn’t discover that particular AI student lark. “The first e-commerce,” Raj Reddy joked at a symposium in 2015.)
In March 1975, after describing my computational life to a skeptical group of high school science and mathematics teachers, I realized how far I’d come from my original agnosticism, how persuaded I was by AI. “I’m no longer a disinterested party,” I confessed to my journal.
In the mid-1970s I lived in a world that would be commonplace three decades later—a computer-saturated environment; email; AIs both visible and invisible, audible and inaudible, and all of them partial intelligences (which threw off outsiders who still held that intelligence was—or it wasn’t). But few others shared such a life, so no wonder outsiders found it difficult to believe on any level. Yet I claimed I was writing science, not science fiction.
Maybe because this world was everyday to me, I couldn’t write Pop! Wow! Zowee! I thought the facts supplied their own pop-wow-zowee without rhetorical flourishes. I hoped to tell the story straightforwardly, with an earnestness that, okay, maybe skirted scholarly tedium. I failed to see how improbable the AI world seemed to anyone outside the field.
My agent and I parted, and I found another who was more enthusiastic. But the rejections kept coming. I was so demoralized after nearly two years of trying to sell the book that Joe offered to intercept the rejections for me. If I had to read these responses, most of them ignorant, not all of them courteous, I wouldn’t have the fortitude to go on. Halfway through the manuscript, I wondered if I should even bother finishing. The work was hard; nobody wanted, much less appreciated it. Years wasted. The Two Cultures clashed again, and as far as New York publishing was concerned, I’d picked the wrong side. They could hardly conceive there was another side to pick.
From my journal, November 28, 1976:
I could use someone to talk to just now, to explain, or figure out, why I feel under such a strain, so dissatisfied, so blue. Partly it has to do with my work. I’m 36, and my work will never be better (I think) yet it’s essentially being ignored. I’m sick to death of the AI project—it takes much longer than I thought—yet wouldn’t dream of not finishing, even though I resent it taking the best of me and my time just now. And even with this sure-fire, can’t-miss project, publishers rise on the horizon like mirages, only to dissolve when I get close. So my professional life seems a shambles, my personal life all screwed up on account of first, my professional life, and second, that I’ve come here to Pittsburgh with Joe on account of his professional life, and I wouldn’t regret that, to speak of, except for his own deep doubts about whether he made the right choice, putting professional before personal.
So I feel like a rat in a maze, with no exit except to grow old. Yet that passivity violates every sense I’ve had of myself since I was 26—ten years now, nearly a quarter of my life. Just writing about it makes me angry at myself, and anxious to grab my life and shake it into shape. But sometimes I feel like one of those Beckett characters: “I can’t go on, I must go on, I can’t go on, I’ll go on.”
I wanted the AI enterprise—theirs and mine—to succeed, though I had only vague ideas of what success would mean and when it might arrive. If the field had seemed mysterious, it was no longer. Difficult, yes. Scientists I interviewed and wrote about worked tirelessly to gain centimeters of progress. From the beginnings of history, humans had imagined creating intelligence outside the human cranium, and later, hoped to understand the human mind in a scientific way. With AI, they were trying to do both.
Finally, after more than thirty rejections (outdoing Catch-22 by several), Joe made a couple of phone calls and put my agent in touch with W. H. Freeman, the book-publishing arm of Scientific American. An editor there, Peter Renz, was intrigued. Renz and I had several meetings, and he seemed enthusiastic.
But between the day Renz said he’d offer me a contract and its arrival, some four months elapsed. Fate intervened melodramatically. In a routine medical visit, a deep internal tumor thought benign had tripled in size within six months. Maybe not benign after all. A second opinion wasn’t reassuring. I’d need surgery to tell for sure. At thirty-six, I came face to face with mortality.
What would you do if you had only a year to live?
A game we’d played as young adults, but now it wasn’t theoretical. I faced the question for real and knew instantly. I must get out of a place that made me miserable and back to warmth, sunshine, beauty, and my family. I had to stop deferring my own desires until the perfect job for Joe came along.
The tumor proved benign after all, but the crisis focused me. Even as I signed the contract with W. H. Freeman—May 22, 1977—and was still recovering from surgery, my path ahead was clear: finish the book and get out of Pittsburgh. It didn’t have to be in that order.
Lois Fowler, a professor of English at Carnegie Mellon, and my closest friend in Pittsburgh; the novelist Mark Harris, my colleague and friend in the English department at Pitt; my friend Herb Simon; my husband Joe; each urged me to stay another year to go through the tenure process. It would be an asset on my resume, they said. My department chair had already encouraged me to start the process, and no one anticipated problems—the Pitt English department’s tenured faculty then comprised scholars who’d published an article or two, or none. I was under contract for a third book. My student evaluations were fine. “All right,” I said, “one more year.”
But the surgery had distressed me at levels I hardly understood. Depression turned into desperation. Doubtful, but knowing I couldn’t go on as I was, I took my friend Lois’s advice and began therapy with a psychiatrist, Charlotte Babcock, who might at least see me through this further purgatorial year.
And I wrote. When I had what I thought was a decent draft, I gave copies to Newell, Simon, and Marvin Minsky to check for technical details. From Newell, I received a four-page single-spaced commentary, very perspicacious, I noted in my journal. He followed that a few days later with more comments, speaking astutely about my persona in the book. On September 28, 1977, I also received two pages of comments from Simon. “They both care very much that this book be done right, and have been monumentally supportive,” I wrote in my journal.
What perturbed Simon was whether I’d got right the impact of the Logic Theorist, the first example of a symbolic thinking machine, and the information-processing model that he and Newell had presented at the Dartmouth meeting. He didn’t think so.
October 7, 1977:
Worked, tussled later with Herb about whether the information-processing model had really been seen as important at the time. We went over Minsky’s early version of “Steps Toward Artificial Intelligence,” and it seems not. Gamely, Herb conceded the point. But I’ll rewrite to show the paradigm shift was heralded at Dartmouth, if not recognized.
October 8, 1977:
Worked the day revising the Dartmouth chapter, and finally got a hypothesis that fits the data and even sounds plausible to Herb, whom I called in the early evening to check it out with. The information-processing model seemed relevant only to psychologists, and most AI types didn’t feel compelled to make AI resemble natural intelligence, necessarily. When I called Herb, he told me he’d spent the day digging around in his own archives, trying to solve the problem. Also, as I was going over the comments he’d made in the margins, I saw the first vulgarity I’ve ever known him to use: SHIT! he’d scrawled at something Lotfi Zadeh had said. I laughed at that, and plenty of other comments he’d made. That marked-up version of my ms will go into the same archives as the tapes and transcripts. What a treasure he is.
October 10, 1977:
Sat down to write a note to Marvin on the ARPANET [early email] and discovered the line-by-line comments from Allen on my first two chapters running on and on and… Again, my heart sank. But they’re not so much arguments as musings, which is even better. And reminders of promissory notes I give the reader to be paid off in due time. I was very, very pleased to get such detailed, loving attention.
Minsky sent email saying “Just lovely.” I wondered if he’d even read it, but later, he commented on the manuscript in detail.
November 4, 1977:
A much greater sense today of being “finished” with the book, a sense of closure, of tying up loose ends. I’m pleased with the last chapter, despite its windiness. I’ve tried to capture the sense of it all being tentative, a prologue to the really big stuff.
I delivered a finished manuscript to W. H. Freeman during the Christmas holidays of 1977, relieved to be finished at last.
But that spring of 1978, the tenure process did not go well. My department recommended me only by a ratio of two-to-one, unhappy with how difficult it was to categorize me. Was I a novelist, as my first two books said, or a nonfiction writer, as this new book seemed to say? The doubters couldn’t understand where I was going as a writer (as if I knew, as if writers methodically laid out a life plan) while the writers on the faculty argued that my commitment was to writing itself.
I wasn’t blameless. In 1976 I’d written a provocative piece that appeared in The Chronicle of Higher Education called (not by me but by the editor) “An Introduction to the Humanities with Dr. Ptolemy.” There I’d argued that the humanities were demoralized, and their exaltation of the human species alone—human chauvinism, both Lewis Thomas and Carl Sagan had called it—was as out-of-date for a guide to being human as the Ptolemaic system was for navigating the high seas.
I referred to a long exchange between C. P. Snow (remember The Two Cultures?) and F. R. Leavis, an influential British critic at Cambridge. Leavis had written that the world’s problems were to be solved by “mankind . . . in full intelligent possession of its full humanity . . . something with the livingness of the deepest vital instinct; as intelligence, a power—rooted, strong in experience, and suprememly human—of creative response to the new challenges of the time. . . ” (Leavis, 2013). I said I’d read this aloud to my students, and asked them what it meant. They laughed. I confessed that I’d laughed with them. “If we’d written that in one of our papers—“ one student began. But Leavis was one of the greatest of the humanists, defending the humanities. “The humanities are demoralized because they are no longer adequate for us in the world as it is,” I wrote. “We are hard up for a Copernican revolution that will take man from the center of the universe and put him somewhere more appropriate.” (McCorduck, 1976)
This offended my humanities colleagues, and the lukewarm tenure vote was to let me know.
How did I feel? “Like I’ve been gangbanged by dwarfs,” I said to somebody at the mailboxes, which went viral around the department. “Like I’ve done right and been deeply wronged,” I wrote in my journal, which went on to note:
In a way it’s laughable, and in a way it’s terrible. I told Mary [the department chair] that what I resent most of all is being penalized for stretching, taking risks; that I’ve worked very hard and stretched far, and it’s not so perceived by my colleagues. I saw Mark Harris briefly. “Don’t take it personally,” he said. “Other things are going on.” But how else should I take it?
Later, a male colleague, here nameless out of courtesy, said:
There’s no way you can prove it; there’s no way you could make a case, but I believe you were discriminated against because you were a very threatening woman to some of the men who cast votes. I picked it up from body language, from the strange nature of the comments that were sometimes made.” I said I’d suspected but hadn’t wanted to believe it. Told Lois, who said Mark said he was absolutely baffled by the comments and the vote. That might account for it.
It was piquant to hear later that this sorrowful male colleague had helped lead the charge against me.
More nonsense leaked out. Some of my colleagues dismissed Herb Simon’s letter on my behalf as merely one of my husband’s “business associates.” Others cried that Machines Who Think showed I’d “sold out to the machines.”
Sold out to the machines? Sloan Wilson, author of that emblematic 1950s novel about selling out, The Man in the Gray Flannel Suit, wrote: “Selling out was doing something you did not want to do for a good deal more money than you got for doing what you loved to do” (Halberstam, 2012).
No. I was doing exactly what I loved and getting nothing but grief. Having gone through the same arguments for two years with publishers, how could I be surprised? I, Simplicissima.
The official notice from the dean said that the university preferred to wait with the tenure decision until they could see how Machines Who Think was received. That struck me as a dismal acknowledgment of intellectual poverty—just what I’d complained about in “An Introduction to the Humanities with Dr. Ptolemy.”
I’ll never know my greater offense: being a strong feminist or being eager to move on from the past, compelled by what I saw as the future. It didn’t really matter. My colleagues had released me. I exhaled. I saw again how far I’d moved from other humanities scholars. The distance across Panther Hollow between the University of Pittsburgh’s English department and Carnegie Mellon’s computer science department yawned as wide and deep as the Two Cultures could possibly be from each other.