John McCarthy (computer scientist)

• 1. Introduction

  ›

  John McCarthy coined the term "artificial intelligence" in the summer of 1956, and spent the rest of his life trying to build exactly what that phrase promised. Born in Boston, Massachusetts, on the 4th of September 1927, to an Irish immigrant father and a Lithuanian Jewish immigrant mother, McCarthy grew up in a family that moved constantly during the Great Depression. His father eventually found work as an organizer for the Amalgamated Clothing Workers in Los Angeles. It was an itinerant, politically charged childhood that shaped a man who would one day ask whether machines could think.

  By the time McCarthy died at his home in Stanford on the 24th of October 2011, he had founded a discipline, invented a programming language that outlasted him, and debated the nature of mind with philosophers who thought he was wrong about almost everything. How did a kid who taught himself university mathematics from borrowed textbooks become one of the architects of modern computing? And what did he actually believe machines could become?

• 2. A Mind Ahead of Schedule

  ›

  McCarthy graduated from Belmont High School two years early and was accepted into Caltech in 1944. He had already taught himself college-level mathematics during his teens by studying Caltech's own textbooks, allowing him to skip the first two years of math once he arrived. A lecture by John von Neumann at Caltech planted the seed that would drive his career.

  His time at Caltech was not entirely smooth. He was suspended for failing to attend physical education courses, served in the US Army, and then returned to complete a Bachelor of Science in mathematics in 1948. He moved on to Princeton University, where he earned a PhD in mathematics in 1951. His dissertation, "Projection operators and partial differential equations", was supervised by Donald C. Spencer. From Princeton he took short-term positions before arriving at Dartmouth College as an assistant professor in 1955, putting him in the room where the defining moment of his field was about to happen.

• 3. Naming a Field

  ›

  In the summer of 1956, a small group of researchers gathered at Dartmouth College for a workshop that changed the course of computing. McCarthy, Marvin Minsky, Nathaniel Rochester, and Claude E. Shannon wrote the proposal for that conference, and in doing so they coined the term "artificial intelligence." It was the moment a scattered set of ideas became a discipline.

  Minsky later joined McCarthy at MIT in 1959, deepening a collaboration that would shape early AI research. McCarthy moved to MIT as a research fellow in the autumn of 1956, where his students affectionately called him "Uncle John" by the end of his time there. In 1962 he became a full professor at Stanford, remaining there until his retirement in 2000. At Stanford he helped establish the Stanford AI Laboratory, which became a friendly rival to Project MAC at MIT, a project he had also helped motivate during his earlier years at MIT.

• 4. Lisp, Garbage, and Time

  ›

  In the late 1950s, McCarthy discovered that primitive recursive functions could be extended to compute with symbolic expressions. The result was Lisp, a functional programming language that also introduced lambda notation borrowed from lambda calculus. Lisp was published in 1960 and quickly became the programming language of choice for AI applications. Later dialects such as Scheme built their semantics on the lambda notation McCarthy introduced.

  Around 1959 he also invented garbage collection, a method of automatic memory management that solved problems arising in Lisp. The concept, now fundamental to countless programming languages, emerged from a practical need rather than a theoretical exercise. That same year he proposed the use of recursion and conditional expressions to an Association for Computing Machinery committee that was designing ALGOL 60, and those ideas were incorporated into the language. McCarthy also became a member of the International Federation for Information Processing Working Group 2.1, the body that specified and maintained ALGOL 60 and ALGOL 68.

  Time-sharing, another of his central contributions, would prove to have consequences he could barely have imagined. His colleague Lester Earnest later told the Los Angeles Times that the internet would not have arrived nearly as soon without McCarthy's push to develop time-sharing systems. What Earnest called "servers" and then "cloud computing" was, in his telling, still just time-sharing. John started it.

• 5. Computing as a Utility

  ›

  In 1961, at a speech celebrating MIT's centennial, McCarthy proposed an idea that sounded almost fantastic at the time: that computer time-sharing technology might one day let computing power, and even specific applications, be sold through a utility business model, the way water or electricity are sold. He was perhaps the first person to suggest this publicly.

  The concept of a computer or information utility was widely discussed during the late 1960s but faded from serious consideration by the mid-1990s. After 2000, it resurfaced in the forms of application service providers, grid computing, and cloud computing. McCarthy was also instrumental in the creation of three of the very earliest time-sharing systems: the Compatible Time-Sharing System, the BBN Time-Sharing System, and the Dartmouth Time-Sharing System. In 1958 he had proposed the advice taker, a program that inspired later work on question-answering and logic programming. And in 1966, his team at Stanford wrote a program to play chess against counterparts in the Soviet Union; the team lost two games and drew two, a series now known as Kotok-McCarthy.

• 6. Can Machines Believe?

  ›

  In 1979, McCarthy published an article titled "Ascribing Mental Qualities to Machines." In it he wrote that machines as simple as thermostats can be said to have beliefs, and that having beliefs seems to be a characteristic of most machines capable of problem-solving performance. The philosopher John Searle responded in 1980 with his famous Chinese Room Argument, disagreeing with McCarthy and arguing that machines cannot have beliefs because they are not conscious. Searle's position was that machines lack intentionality.

  In a 1989 television interview with Jeffrey Mishlove for ThinkingAllowedTV, McCarthy addressed the question of whether intuition separates humans from computers. He acknowledged that some people believe humans possess a spiritual or transcendent intuition beyond physical reach, but he noted that such a view had been in steady decline for a few centuries. He expected the challenges of realizing human consciousness in machines to be overcome.

  His most persistent intellectual opponent was Hubert Dreyfus, a philosophy professor at Berkeley University. Dreyfus saw human reasoning as something deeper than logic, rooted in existential questions about the good life and nihilism. McCarthy, by contrast, was confident that every aspect of human intelligence could be formalized precisely enough to be programmed into a machine. Their disagreement ran the length of both their careers, with McCarthy remaining outwardly optimistic to the end. From 1978 to 1986 he developed the circumscription method of non-monotonic reasoning, an approach to making machines capable of common-sense inference under incomplete information.

• 7. The Man Behind the Field

  ›

  McCarthy's Usenet signature block carried a message he held for years: "He who refuses to do arithmetic is doomed to talk nonsense." His license plate cover said essentially the same thing. He advised 30 PhD graduates over the course of his career and was an avid reader, a staunch supporter of free speech, and a regular at San Francisco Bay Area dinners in Palo Alto attended by readers of the rec.arts.books newsgroup, events they called rab-fests.

  His political journey was as unusual as his intellectual one. Both of his parents were active members of the Communist Party during the 1930s, and they encouraged critical thinking and learning. McCarthy became a conservative Republican after visiting Czechoslovakia in 1968, following the Soviet invasion. His mother had died in 1957. His second wife, Vera Watson, was a programmer and mountaineer who died in 1978 while attempting to scale Annapurna I Central as part of an all-women expedition. He later married Carolyn Talcott, a computer scientist at Stanford and later at SRI International.

  In 2001 he published a short story, "The Robot and the Baby," which farcically explored whether robots should have or simulate emotions, and which anticipated aspects of internet culture and social networking. The awards that marked his career included the 1971 Turing Award for his contributions to AI, the Kyoto Prize in 1988, the National Medal of Science in 1990, and the Benjamin Franklin Medal in Computer and Cognitive Science in 2003. In 1982 he also appears to have originated the idea of the space fountain, a tower extending into space kept upright by a stream of pellets propelled from Earth, with payloads riding a conveyor belt upward, a concept that pointed his restless imagination far beyond the machines he had already built.