Warren Sturgis McCulloch (1898-1969) was an American neurophysiologist and cybernetician whose lifelong question was the relationship between the physiology of the brain and the nature of thought. Trained in philosophy, psychology, and medicine, he held positions at Yale and the University of Illinois College of Medicine before joining the Research Laboratory of Electronics at MIT in the early 1950s, where he remained for the rest of his career. He was a leading force in establishing cybernetics, the study of control and communication in animals and machines, and chaired the influential Macy Conferences that gave the field its shape.
McCulloch’s most lasting contribution is the 1943 paper “A Logical Calculus of the Ideas Immanent in Nervous Activity,” written with the young self-taught logician Walter Pitts. It treated each neuron as a simple all-or-none switch and proved that networks of such units could compute any proposition of logic. This was the first attempt to describe the neuron in precise mathematical terms, and it built a bridge between the biology of the brain and the formal logic underlying computing machines.
For the library’s reader, McCulloch sits at the headwaters of an idea that now dominates AI. The McCulloch-Pitts neuron is the direct conceptual ancestor of the artificial neurons in every modern neural network. The honest limit, worth keeping in view, is that his model was about computation, not learning: a McCulloch-Pitts network has fixed wiring and cannot adjust itself from experience. Supplying that missing piece was left to others, from Hebb’s learning rule in 1949 to the perceptron and, much later, backpropagation.