Walter Pitts

• 1. Introduction

  ›

  Walter Pitts was born in Detroit, Michigan, on the 23rd of April 1923, and he died at age 46, leaving behind a body of work that still defines how scientists think about the brain as a computing device. He never earned a doctorate. He spent years without a home. He once burned his own dissertation. Yet the model he proposed with Warren McCulloch in 1943 remains the standard reference in the entire field of neural networks, and the unit at its center still carries both their names.

  How does a self-taught child from Detroit, who spent three days in a library reading Bertrand Russell at age 12, end up reshaping neuroscience, computer science, cognitive psychology, and artificial intelligence? What drove him to erase himself from public record, refuse every advanced degree offered to him, and ultimately destroy years of his own work? Those are the questions at the heart of this story.

• 2. A Child Who Read Russell

  ›

  At age 12, Walter Pitts walked into a library and spent three days reading Bertrand Russell's Principia Mathematica. According to his biographer, Neil R. Smalheiser, that experience triggered a metaphysical conviction in Pitts: logic ruled the universe, and personal ego was an obstacle to understanding it. That belief would shape everything he did afterward, including his refusal to sign his name to anything official.

  Pitts wrote a letter to Russell pointing out what he considered serious problems with the first volume's first half. Russell, far from dismissing the letter, was appreciative and invited Pitts to study at Cambridge University. Pitts did not take up that offer. Instead, at age 15, he left home and went to Chicago, where Russell was delivering lectures at the University of Chicago in the fall of 1938. Pitts attended without ever registering as a student. A friend, Jerome Lettvin, later described an anecdote, possibly apocryphal, in which Pitts met Russell for the first time in Chicago, held a conversation without realizing who Russell was, and still managed to impress him.

• 3. The Chicago Years

  ›

  In 1938 Pitts met Jerome Lettvin, then a pre-medical student, and the two became close friends. Russell directed Pitts to study with the logician Rudolf Carnap, and Pitts introduced himself to Carnap in a characteristically unconventional way: he walked into Carnap's office during office hours and handed over an annotated copy of Carnap's own recent book, The Logical Syntax of Language, without giving his name. Carnap spent months searching for the young annotator before finding him.

  Once Carnap located Pitts, he secured him a menial job at the university. Pitts at that time was homeless and without income. He mastered Carnap's abstract logic and then became drawn to Nicolas Rashevsky, a Ukrainian mathematical physicist at Chicago who founded the field of mathematical biophysics by recasting biology in the language of the physical sciences and mathematical logic. Pitts attended Rashevsky's seminars in theoretical biology regularly, alongside figures including Frank Offner, Herbert Landahl, Alston Scott Householder, and the neuroanatomist Gerhardt von Bonin from the University of Illinois at Chicago.

  In 1940, von Bonin introduced Lettvin to Warren McCulloch, who would eventually become a professor of psychiatry at Illinois. That introduction set the stage for one of the most consequential collaborations in the history of brain science.

• 4. The 1943 Paper

  ›

  In early 1942, Warren McCulloch invited Pitts, still homeless, and Lettvin to live with his family. The evenings became working sessions. Pitts brought his knowledge of Gottfried Leibniz's work on computing, and together he and McCulloch examined whether the nervous system could function as a kind of universal computing device in the sense Leibniz had described. Out of those conversations came "A Logical Calculus of Ideas Immanent in Nervous Activity," published in 1943.

  The paper proposed the first mathematical model of a neural network. The formalized neuron at its center, now called the McCulloch-Pitts neuron, remains the standard reference in the neural networks field more than eight decades later. The University of Chicago, acknowledging the significance of Pitts's contribution after five years of unofficial study, awarded him an Associate of Arts, the only earned degree he ever held.

  Before that landmark paper, Pitts had already published his ideas about building a Turing machine in The Bulletin of Mathematical Biophysics, in an essay titled "Some Observations on the Simple Neuron Circuit."

• 5. MIT and Norbert Wiener

  ›

  In 1943, Lettvin introduced Pitts to Norbert Wiener at MIT. Their first conversation centered on Wiener's proof of the ergodic theorem, and it went so well that Pitts moved to Greater Boston to work alongside him. He formally enrolled in MIT's physics department during the 1943-1944 academic year and later in the electrical engineering department from 1956 to 1958, even as he functioned in practice as an unofficial student under Wiener's direction.

  From 1946 onward, Pitts was a core member of the Macy conferences, gatherings whose stated purpose was to establish the foundations for a general science of how the human mind works. By 1951, Wiener had convinced Jerome Wiesner to hire physiologists of the nervous system, and a group formed around Pitts, Lettvin, McCulloch, and Pat Wall. Lettvin described Pitts during this period as "in no uncertain sense the genius of the group," adding that a question put to Pitts would return an entire textbook in reply.

  Pitts was writing a large dissertation on the properties of neural nets connected in three dimensions. He also refused all offers of advanced degrees and positions of authority at MIT, partly because accepting would have required him to sign his name publicly.

• 6. The Break and the Burning

  ›

  In 1952, Norbert Wiener's wife, Margaret Wiener, turned sharply against McCulloch, and Wiener abruptly severed relations with everyone connected to him, including Pitts. The rupture left Pitts increasingly isolated. He remained employed as a research associate in MIT's Research Laboratory of Electronics, but the position became, as sources described it, little more than a technicality.

  Then in 1959 came the publication that broke what remained of his scholarly momentum. "What the Frog's Eye Tells the Frog's Brain," credited to Humberto Maturana, Lettvin, McCulloch, and Pitts, demonstrated that analog processes in the eye were doing at least part of the interpretive work in image processing, rather than the brain computing information digital neuron by digital neuron through mathematical logic alone. For Pitts, this finding directly contradicted the logical-digital framework he had built his career upon. He burned his unpublished doctoral dissertation on probabilistic three-dimensional neural networks and destroyed years of unpublished research.

  His last published collaboration was a 1965 paper on olfaction with Lettvin and Robert Gesteland, titled "Chemical Transmission in the Nose of the Frog."

• 7. Legacy

  ›

  Walter Pitts died on the 14th of May 1969 of bleeding esophageal varices, a condition typically associated with cirrhosis and alcoholism. He was 46. He never married. He spent much of his adult life without a permanent home and refused public recognition at every turn.

  The field he helped create did not wait for his permission to grow. The McCulloch-Pitts neuron became the conceptual foundation for artificial neural networks, and his 1943 paper influenced cognitive sciences, psychology, neuroscience, computer science, cybernetics, and artificial intelligence. McCulloch and Pitts also published a 1947 paper on the perception of auditory and visual forms, extending their theoretical reach into how the brain recognizes patterns. The Macy conferences, in which Pitts was a central figure from 1946 onward, planted the seed for cybernetics as a discipline. The intellectual lineage running from those evenings at McCulloch's house to modern machine learning passes directly through the simple formalized neuron that a homeless teenager helped invent.