Hubert Lederer Dreyfus and Herbert Paul Grice, Philosophers of the Limits of Computers
Hubert L. Dreyfus in 1999. His 1972 book “What Computers Can’t Do” made him a scourge and eventually an inspiration to researchers in artificial intelligence. Michael J. Okoniewski
Hubert L. Dreyfus, a philosopher whose 1972 book “What Computers Can’t Do” made him a scourge and eventually an inspiration to researchers in artificial intelligence, died on April 22 at his home in Berkeley, Calif. He was 87.
The University of California, Berkeley, where he was a longtime professor of philosophy, said the cause was cancer.
Professor Dreyfus became interested in artificial intelligence in the late 1950s, when he began teaching at the Massachusetts Institute of Technology. He often brushed shoulders with scientists trying to turn computers into reasoning machines.
“They said they could program computers to be intelligent like people,” he recalled in a 2005 interview with the blog Full-Tilt Boogie. “They came to my course and said, more or less: ‘We don’t need Plato and Kant and Descartes anymore. That was all just talk. We’re empirical. We’re going to actually do it.’”
He added: “I really wanted to know, could they do it? If they could, it was very important. If they couldn’t, then human beings were different than machines, and that was very important.”
In 1965, after spending time at the RAND Corporation, he published “Alchemy and Artificial Intelligence,” a slashing attack on the work of Allan Newell and Herbert A. Simon, two of RAND’s leading artificial intelligence researchers, and followed with the equally provocative “What Computers Can’t Do: A Critique of Artificial Reason.”
Professor Dreyfus argued that the dream of artificial intelligence rested on several flawed assumptions, chief among them the idea that the brain is analogous to computer hardware and the mind to computer software.
In this view, human beings develop an accurate picture of the world by adding bits of information and rearranging them in a procedure that follows predictable rules.
Professor Dreyfus, an adherent of the French phenomenologist Maurice Merleau-Ponty and the German philosopher Martin Heidegger (he had written seminal introductory works on both men), posited a different view of human beings and their interactions with the world around them.
There was no objective set of facts outside the human mind, he insisted. Human beings experienced learning as a partly physical interaction with their surroundings, and interpreted the world, in a process of continual revision, through a socially determined filter.
Inevitably, he said, artificial intelligence ran up against something called the common-knowledge problem: the vast repository of facts and information that ordinary people possess as though by inheritance, and can draw on to make inferences and navigate their way through the world.
“Current claims and hopes for progress in models for making computers intelligent are like the belief that someone climbing a tree is making progress toward reaching the moon,” he wrote in “Mind Over Machine: The Power of Human Intuition and Expertise in the Era of the Computer” (1985), a book he collaborated on with his younger brother Stuart, a professor of industrial engineering at Berkeley.
His criticisms were greeted with intense hostility in the world of artificial intelligence researchers, who remained confident that success lay within reach as computers grew more powerful.
When that did not happen, Professor Dreyfus found himself vindicated, doubly so when research in the field began incorporating his arguments, expanded upon in a second edition of “What Computers Can’t Do” in 1979 and “What Computers Still Can’t Do” in 1992.
Hubert Lederer Dreyfus, known as Bert, was born on Oct. 15, 1929, in Terre Haute, Ind. His father, Stanley, was in the wholesale poultry business, and his mother, the former Irene Lederer, was a homemaker.
At Wiley High School, his debate coach encouraged him to apply to Harvard, which he thought was in England, because of the Cambridge address. He was more interested in a different Cambridge school that he thought might sharpen his talent for making homemade explosives and setting them off by remote control.
“I wanted to go to M.I.T. because I figured they would help me make better bombs,” he said in an interview at the Institute of International Studies at Berkeley in 2005.
In the end, he opted for Harvard, where he studied physics initially but switched majors after hearing a lecture by the American philosopher C. I. Lewis.
He received a bachelor’s degree in philosophy in 1951, writing an undergraduate thesis on causality in quantum mechanics, and a master’s degree in 1952. Before completing his doctorate in 1964, with a dissertation on Edmund Husserl — a philosopher he later dismissed as “boring” — he spent fellowships in Freiburg, Germany; Louvain, Belgium; and the École Normale Supérieure in Paris, absorbing the latest developments in Continental philosophy.
After returning to the United States, he taught at Brandeis University and M.I.T. and translated, with his first wife, the former Patricia Allen, Merleau-Ponty’s “Sense and Non-Sense,” published in 1964. He joined the philosophy department at Berkeley in 1968.
His first marriage ended in divorce. In addition to his younger brother, Stuart, an emeritus professor of industrial engineering at Berkeley, he is survived by his wife, Geneviève Boissier-Dreyfus, and their two children, Stéphane and Gabrielle Dreyfus.
Professor Dreyfus went on to play a major role in explaining Continental thought in works like “Michel Foucault: Beyond Structuralism and Hermeneutics” (1982), written with Paul Rabinow, and “Being-in-the-World: A Commentary on Heidegger’s ‘Being and Time, Division I’” (1989). With Mark Wrathall, a professor of philosophy at the University of California, Riverside, he edited numerous guides devoted to existentialism, phenomenology and Heidegger’s philosophy.
“It is no exaggeration to say that, insofar as English speaking philosophers have any access at all to thinkers like Heidegger, Merleau-Ponty and Michel Foucault, it is through the interpretation that Dreyfus originally offered of them,” the Harvard philosophy professor Sean D. Kelly wrote recently on the philosophy website Daily Nous.
In later years, he turned his attention to new subjects. With Professor Kelly, he wrote a surprise best seller on literature, “All Things Shining: Reading the Western Classics to Find Meaning in a Secular Age” (2011). In “Skillful Coping: Essays on the Everyday Phenomenology of Everyday Perception and Action” (2014), an essay collection edited by Professor Wrathall, he employed the insights of phenomenology to explore nonreflexive action and ethics.
For his 2006 book “Philosophy: The Latest Answers to the Oldest Questions,” Nicholas Fearn broached the topic of artificial intelligence in an interview with Professor Dreyfus, who told him: “I don’t think about computers anymore. I figure I won and it’s over: They’ve given up.”